Objective: This study analyzes whether small-diameter Contegras behave in the same way as small-diameter homografts, when implanted for the first time in pulmonary position. Methods: Small-diameter conduits include 12 and 14 mm Contegras and 8-14 mm homografts. Graft dysfunction is defined as right ventricular outflow tract obstruction with peak echo-Doppler gradient > 40 mmHg, or grade III/IV graft regurgitation. Graft failure is defined as need for conduit replacement or need for catheter or surgical reintervention. Thirty-eight patients who received small Contegras (n = 25) and small homografts (n = 13) from October 2002 to end December 2006 were studied. The most frequent indication was pulmonary atresia and ventricular septal defect (n = 20; 10 associated with major aorto-pulmonary collateral arteries), followed by truncus arteriosus (n = 12). Most patients' characteristics were comparable except that recipients of homografts were smaller (p for body area = 0.014). Survival, freedom from graft dysfunction, failure and explantation were estimated by the Kaptan-Meier method. The log-rank test was used to compare outcomes. Results: There were three early and four late deaths. No death was graft related. Survival was 80 +/- 8.2% for patients with Contegras and 77 +/- 11.7% for those with allografts: p = 0.82. Mean follow-up duration is 22 16 months. Freedom from dysfunction for Contegra conduits decreased in the first 6 months and stabilized at 58 +/- 11% from month 14. For homografts it decreased only 1 year after implantation, down to 35 +/- 19.7% from month 31: p = 0.61. Freedom from Contegra failure diminished the first 16 months to level out at 57 +/- 13%. No homograft failed the first 2 years. With a p-value of 0.14, homografts tended to fail less frequently. Five grafts were exptanted. Freedom from explantation was similar (p = 0.98): 90 +/- 6.7% for Contegras and 75 +/- 21.6% for homografts at year 3. Conclusion: In the first 4 years after pulmonary implantation of small-diameter Contegras and homografts, the fate of both conduits was statistically similar, in spite of different behavior. As Contegra is 'off-the-shelf' available, it constitutes a sound alternative to homograft for right ventricular outflow tract reconstruction in neonates and infants. (C) 2007 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved.

Nachfolgend werden 2 Fälle nekrotisierender Pneumonie im Kindesalter (18 Monate und 4 Jahre) vorgestellt, die eine Lobektomie bzw. Pneumonektomie erforderten. Als Erreger wurden beim ersten Patienten Bacillus cereus und Aspergillus fumigatus isoliert, die eine postoperative nosokomiale Infektion nach totaler kavo-pulmonaler Konnektion verursachten. Beim 2. Patienten wurden im Trachealsekret α-hämolysierende Streptokokken als Ursache eines resistenten parapneumonischen Empyems nachgewiesen. Beide Patienten sind schließlich genesen. In den hochentwickelten Industrienationen treten selten derart komplizierte Pneumonien im Kindesalter auf, die eine Lungenresektion erforderlich machen. Nur eine ausgeprägte Lungennekrose indiziert solche Maßnahmen, und der Zeitpunkt der chirurgischen Intervention muss sensibel gewählt werden, um einen günstigen Verlauf zu erzielen.

In this paper two cases of necrotizing pneumonia, one of which required lobectomy and one, pneumonectomy, in children aged 18 months and 4 years, respectively, are presented. In the former patient the pathogens concerned were Bacillus cereus and Aspergillus fumigatus, which had led to postoperative nosocomial infections after total cavopulmonary connection. Alpha hemolytic streptococci were found in the tracheal secretion in the second patient, which had caused treatment-resistant parapneumonic empyema. Both patients eventually recovered. Complicated pneumonia requiring lung resection is rarely encountered in children in industrialized countries. Extensive pulmonary necrosis is the only indication for such surgery. The surgical operation must be carefully timed for a favourable outcome to be achieved.

Objective: Imbalances of pulmonary to systemic blood flow ratio (Q(p)/Q(s)) compounded with inadequate systemic oxygen delivery correlate with mortality after first-stage Norwood palliation of hypoplastic left heart syndrome. Mathematical models suggest that maximal systemic oxygen delivery occurs with Q(p)/Q(s) of less than 1. Whether this applies to clinical practice is unclear. This study evaluates the level of Q(p)/Q(s) that correlates with best hemodynamic status in the first 48 postoperative hours. Methods: Hemodynamic data of 25 consecutive patients who underwent Norwood procedure from October 2002 to January 2005 were retrospectively analyzed. Data included, in particular, systemic venous and arterial oxygen saturation (SvO(2) and SaO(2), respectively), Q(p)/Q(s), lactate levels, and doses of required inotropes. Parameters were recorded 3 hourly. Data were assigned to three groups according to their corresponding Q(p)/Q(s) Groups 1, 2, and 3 for Q(p)/Q(s) <= 1, Q(p)/Q(s) between 1 and 2, and Q(p)/Q(s) : 2, respectively. Thereafter, independent t-test or Fisher's exact test was used to reveal significant differences. Q(p)/Q(s) ratios and lactate levels were compared in hospital survivors and non-survivors. Results: Out of 343 samples, 110, 184, and 49 were assigned to groups 1, 2, and 3, respectively. Group 1 (Q(p)/Q(s) < 1) was characterized by lower SaO(2) (p < 0.001) with similar SvO(2) (p = 0.3 and p = 0.5) and, therefore, higher systemic oxygen delivery (arteriovenous oxygen saturation difference, p < 0.001; oxygen excess factor, p < 0.001) compared to groups 2 and 3. However, tower mean arterial pressure (p = 0.07 and p < 0.001), higher lactate levels (p = 0.009 and p = 0.01), and norepinephrine doses (p = 0.006 and p < 0.001) highlighted worse hemodynamics. The best hemodynamic status corresponded to group 2. Q(p)/Q(s) remained above 1 in 21 survivors and was, most of the times, below 1 in four patients who died. Lactate levels were almost always above 4 mmol/l or increasing in non-survivors. Conclusions: Maximum oxygen delivery after Norwood operation occurs at Q(p)/Q(s) of less than 1. However, optima( hemodynamic status and endorgan function and higher survival correlates with Q(p)/Q(s) between 1 and 2. Thus, Q(p)/Q(s) should be targeted at 1. 5 for improved course early after first-stage Norwood palliation. (c) 2006 Elsevier B.V. All rights reserved.

Background. Recently introduced cardiopulmonary bypass techniques to avoid circulatory arrest were proposed to improve organ function of the modified Norwood operation for hypoplastic left heart syndrome. This study compares postoperative hemodynamics and survival in patients who underwent Norwood procedure on the beating heart to those operated on with cardioplegic cardiac arrest.Methods. Between October 2002 and January 2005, 26 consecutive patients aged 4 to 275 days (median, 9 days) and weighing 2.9 to 4.4 kg (median, 3.4 kg) underwent Norwood palliation: 13 with continuous coronary and systemic perfusion (group 1), and 13 with only continuous systemic perfusion but arrested heart (group 2). Standard hemodynamic measurements, lactate levels, arterial and superior vena cava oxygen saturations, and inotropic agents required for postoperative hours 0, 6, 12, 18, 24, and 48 were retrospectively analyzed. For univariate comparison of different variables, chi(2) test, Fisher's exact test, or Student's t test was used as appropriate.Results. In group I significantly higher mean arterial pressure (53 +/- 0.8 versus 50 +/- 1.2 mm Hg; p = 0.04), higher central oxygen saturation (54% +/- 1.1% versus 50% +/- 1.5%; p = 0.03), higher urinary output (5.3 +/- 0.4 versus 4.4 +/- 0.4 mL (.) kg(-1) (.) h(-1); p = 0.09), lower lactate levels (2.4 +/- 0.1 versus 4.1 +/- 0.6 mmol/L; p = 0.009) with lower doses of norepinephrine (0.03 +/- 0.004 versus 0.14 +/- 0.03 mu g (.) kg(-1) (.) min(-1); p = 0.002) were recognized. Hospital mortality was 0% in group 1 and 38.5% (5 of 13) in group 2 (P = 0.04). Univariate analysis revealed mortality to be also correlated with preoperative intubation (p = 0.02) and the use of preoperative inotropic agents (p = 0.03).Conclusions. Avoidance of cardiac arrest by means of continuous coronary perfusion in addition to continuous systemic perfusion significantly improves postoperative hemodynamic performance and thus helps to reduce hospital mortality after the modified Norwood procedure.

Objective: To examine early and long-term results of surgical aortic valvotomy in neonates and infants aged less than 3 months. Methods: A review of all 34 neonates (n = 26) and young infants (n = 8) aged 1-62 days undergoing primary open valvotomy for aortic valve stenosis between 1983 and 2003 was carried out. Associated major cardiac anomalies were endocardial fibroetastasis (n = 8), aortic coarctation (n = 3), subvalular aortic stenosis (n = 2), and ventricular septal defect (n = 1). Risk factors for early mortality were estimated. Current information was available for 31 patients for a follow-up of 115 67 months. Kaplan-Meier method was used to estimate freedom from reintervention. Results: Two neonates died early: operative mortality of 6% (2/34). Risk factors for early mortality were associated endocardial fibroetastosis, monocuspid aortic valve and impaired left ventricular function. No patient died late. Seven patients needed reintervention for re-aortic stenosis (n = 5) or aortic insufficiency (n = 2), i.e., re-valvotomy (n = 3), valve replacement (n = 2), Ross procedure (n = 1), and balloon valvuloplasty (n = 1). Freedom from reintervention was 85.1 +/- 6.9%, 78.0 +/- 9.35%, and 53.5 +/- 15.9% at 5, 10, and 15 years, respectively. Conclusions: Primary surgical aortic valvotomy in early infancy carries a tow early and late mortality, a tow occurrence of significant aortic regurgitation and a tow early recurrence of aortic stenosis. In great majority of cases, reintervention can be delayed to allow implantation of an adult-sized prosthesis, when required. (c) 2006 Elsevier B.V. All rights reserved.

Background. Aristotle score is emerging as a reliable tool to measure surgical performance. We estimated the comprehensive Aristotle score for the Norwood procedure, correlated it with survival, and considered its impact on surgical management of hypoplastic left heart syndrome.Methods. Comprehensive Aristotle score was retrospectively calculated for 39 consecutive Norwood procedures performed from 2001 to 2004. Survival was estimated by the Kaplan-Meier method.Results. The Aristotle scores ranged from 14.5 to 23.5 (mean, 19.12 +/- 2.52; median, 19.5). The score was 20 or greater in 44% (17 of 39) of cases. The most frequent patient-adjusted factors were aortic atresia (n = 16), interrupted aortic arch (n = 9), mechanical ventilation to treat cardiorespiratory failure (n = 19) and shock resolved at time of surgery (n = 13). Hospital mortality was 58.8% (10 of 17) in case of score of 20 or more and 9.1% (2 of 22) for score less than 20 (p = 0.0014). From 2003 on, all patients with a score less than 20 survived. Actuarial estimate of survival at 1 year is 56.2% +/- 7.9% and there have been no late deaths after 1 year. One-year survival is much lower (p = 0.001) for patients with scores of 20 or greater (29.4% +/- 11.05%) compared with those whose scores were less than 20 (77.3% +/- 8.9%).Conclusions. This study shows significant correlation of comprehensive Aristotle score with hospital mortality and late survival after Norwood palliation. It suggests that operative survival on the order of 90% may be achieved in patients with comprehensive complexity scores of less than 20. Efforts should be devoted to improve survival of high-risk patients (score >= 20).

Background. The excellent results for repair of truncus arteriosus reported in some centers have not applied to patients with associated interrupted aortic arch. This work aims at understanding the discrepancy of results in our own experience.Patients and Methods. Ten patients among 83 consecutive children with truncus arteriosus repaired from 1987 to September 2004 who had aortic arch interruption were analyzed, with particular emphasis on clinical presentation and outcome. The comprehensive Aristotle complexity score was calculated for each patient. The Kaplan-Meier method was used to estimate survivals.Results. Preoperative mechanical ventilation was necessary in 5 of the 10 patients; 2 of them were moribund. Associated major lesions were as follows: severe (n = 2) and moderate (n = 4) truncal valve regurgitation, coronary artery anomalies (n = 3) and Di-George's syndrome (n = 4). The comprehensive Aristotle score was at least 20 in 6 patients. There were 5 operative deaths (5 of 10); early mortality was 50% (95% confidence limits: 19% to 81%). These deaths occurred in patients with Aristotle score of 20 or greater (5 of 6 = 83%). All 4 patients who had no moderate or severe truncal valve regurgitation survived the intervention. Survival was a low 37.5% +/- 16.1% from 1 year on compared with a high 95.5% +/- 2.5% for the 73 patients without aortic arch interruption.Conclusions. This study confirms the predictive value of the Aristotle score, hospital mortality being significantly correlated with the highest Aristotle score (p = 0.024). To improve outcome in these high-risk patients, preoperative management should be optimized, repair should not be delayed, and regurgitant truncal valve should be repaired or replaced. (c) 2005 by The Society of Thoracic Surgeons.

Objective: Low cardiac output syndrome (LCO) and imbalance of pulmonary and systemic blood flow (Qp/Qs) are common after modified Norwood operation (mNO). Clinical studies indicate that arterial saturations do not correlate with Qp/Qs nor with situations of LCO. Correlation of central venous saturations with Qp/Qs and LCO was evaluated in this study.

The development of aortic regurgitation (AI) is a rare but serious complication of subaortic ventricular septal defects (VSD). Over a period of 5 years we observed VSD-related AI in 24 patients, a frequency of 4.5% of allisolated VSD's encountered during that time frame. The location of the defects was in the infundibular septum in 59%, it was perimembranous in 25% and in the trabecular septum in 16%. Hemodynamically the defects were small except for 2 where the Q p /Q s ratio was > 2. Of the 24 pts, 16 had surgical closure of their VSD accompanied in 9 by aortic valvuloplasty. AI was caused by elongation or defect of the right coronary leaflet in 42%, of the noncoronary leaflet in 25% and a combination of both, in 8%. In 6 pts with infundibular VSD absence of part of the aortic valve ring above the defect was the underlying mechanism for AI. Postoperatively AI was improved to moderate in one pt and to none to trivial in 15. LV end-diastolic diameter decreased significantly in all pts operated. Pathogenetic mechanisms for the development of AI are a deficiency in the aorto-infundibular junction with prolaps of the right -or non-coronary leaflet, deficiency of the valve supporting structures including the valve ring as well as suction of the already elongated leaflet into the VSD with further damage to the antiregurgitant mechanism of the semilunar valve at risk. In perimembranous VSD's, late AI is probably related to turbulent flow through the adjacent LVOT. Surgical closure of isolated VSD's with a location immediately beneath the aortic valve is indicated regardless of their size to prevent the development of AI. If AI has occurred, VSD closure including aortic valvuloplasty improves the amount of regurgitation and normalizes LV enddiastolic dimension.

Das Auftreten einer zusätzlichen Aorteninsuffizienz (AI)ist eine schwerwiegende Komplikation der Ventrikelseptumdefekte(VSD), die wir in einem 5-Jahreszeitraum bei 24 Patienten (17männlich, 7 weiblich) in einer Häufigkeit von 4,5% allerisolierter VSD’s beobachteten. Bei den VSD’s handelte es sich in59% um infundibuläre VSD’s, in 25% um perimembranöse und in 16%um trabekuläre. Hämodynamisch waren die Defekte klein, nur 2Patienten hatten ein Qp/Qs-Verhältnis > 2/1.

Objectives: The use of valved conduits to reconstruct the right ventricular outflow tract for early repair of truncus arteriosus has improved patients'survival. However the burden of cardiological and surgical reinterventions becomes an increasing concern as the child grows.

The development of aortic regurgitation (AI) is a rare but serious complication of subaortic ventricular septal defects (VSD). Over a period of 5 years we observed VSD-related AI in 24 patients, a frequency of 4.5% of all isolated VSD's encountered during that time frame. The location of the defects was in the infundibular septum in 59%, it was perimembranous in 25% and in the trabecular septum in 16%. Hemodynamically the defects were small except for 2 where the Q(p)/Q(s) ratio was > 2. Of the 24 pts, 16 had surgical closure of their VSD accompanied in 9 by aortic valvuloplasty. AI was caused by elongation or defect of the right coronary leaflet in 42%, of the noncoronary leaflet in 25% and a combination of both, in 8%. In 6 pts with infundibular VSD absence of part of the aortic valve ring above the defect was the underlying mechanism for AI. Postoperatively AI was improved to moderate in one pt and to none to trivial in 15. LV end-diastolic diameter decreased significantly in all pts operated. Pathogenetic mechanisms for the development of AI are a deficiency in the aorto-infundibular junction with prolaps of the right-or non-coronary leaflet, deficiency of the valve supporting structures including the valve ring as well as suction of the already elongated leaflet into the VSD with further damage to the antiregurgitant mechanism of the semilunar valve at risk. In perimembranous VSD's, late AI is probably related to turbulent flow through the adjacent LVOT. Surgical closure of isolated VSD's with a location immediately beneath the aortic valve is indicated regardless of their size to prevent the development of AI. If AI has occurred, VSD closure including aortic valvuloplasty improves the amount of regurgitation and normalizes LV enddiastolic dimension.

BACKGROUND:This study evaluates the results of the arterial switch operation for early total repair of double-outlet right ventricle with subpulmonary ventricular septal defect (the Taussig-Bing heart).METHODS:From 1986 through April 2003, 27 patients with Taussig-Bing anomaly underwent arterial switch operation. Twenty patients were neonates (n = 11) or infants younger than 3 months (n = 9). Obstruction of aortic arch (n = 19) or subaortic right ventricular outflow tract obstruction (n = 20) and unusual coronary artery patterns (n = 19) were common. Total correction as a single procedure was performed in 21 patients. Events are depicted by Kaplan-Meier curves.RESULTS:There was 1 patient hospital death at 2 months after repair. One patient died late that was not cardiac related. Survival was 92% +/- 6% at 8 months and remained constant thereafter. Four patients underwent reoperation (1 for residual aortic arch obstruction and 3 for subvalvular and valvular pulmonary stenosis). Freedom from reoperation decreased to stabilize at 83% +/- 8% after 2 years. The risk to have right ventricular outflow tract obstruction develop was 33% +/- 10% at 1 year, increasing slowly and leveling out at 57% +/- 12% at year 5 and thereafter. Statistical analysis revealed no significant risk factor for death or need for reoperation.CONCLUSIONS:The Taussig-Bing anomaly should be corrected in the neonatal period or in early infancy by arterial switch operation, closure of the ventricular septal defect, and simultaneous correction of associated cardiovascular anomalies as a one-stage procedure. Right ventricular outflow tract obstruction often complicates the postoperative course and is the main cause for reintervention.

Background. Limited durability is expected for small homograft valves that are used to correct congenital cardiac disease.

Objectives: To identify potential risk factors influencing early and late outcome following the arterial switch operation (ASO) for transposition of the great arteries associated with ventricular septal defect including double-outlet right or left ventricle. Methods: All patients who underwent ASO in our department until August 2000 (n = 105) were included in this study. There were 77 transpositions of the great arteries with ventricular septal defect. 22 Taussig-Bing hearts and six patients with double-outlet morphology. The median age at operation was 24 days. Aortic arch obstruction was present in 25 patients: in 13 of these patients, a repair with aortic arch reconstruction was done before ASO. The usual coronary artery pattern was present in 59% of the patients. In six patients, we found an intramural course of at least one coronary artery. The ventricular septal defect was closed with a patch through the right atrium (n = 35), the aorta (n = 25), the pulmonary artery (n = 25) or the right ventricle (n = 3); in 17 patients a combined approach was necessary. Results: There were five hospital deaths (4.7%, 95% confidence limit 2-11%). The median duration of follow-up was 72 months. Fourteen patients underwent 15 reoperations 33 months after repair (median), eight for right ventricular outflow tract obstruction or neopulmonary stenosis. Four late deaths occurred, two due to complications related to coronary artery anomalies. Statistical analysis revealed no significant risk factor whatsoever correlating with death or need for reoperation. Survival after 12 years was 91.6%, and freedom from reoperation was 82.6%, Latest follow-up data showed that 13% of patients were in NYHA class II and/or required medical treatment; 87% were in NYHA class I. Conclusions: ASO associated with patch closure of ventricular septal defect can be per-formed early in life with a low risk of mortality (<5%), low incidence of reintervention (< 15%) and promising long-term outcome. (C) 2001 Published by Elsevier Science B.V.

Congenital Heart disease with a poor prognosis has to be operated early but with an acceptable surgical risk and a good chance of survival. The aim of this study was to analyse the indications, the operative mortality and medium-term survival of neonates undergoing cardiac surgery under cardiopulmonary bypass from 1991 to 1998. Three hundred and twenty nine operations were programmed in 326 neonates, 18% (329/1805) of all open heart surgical procedures. Anatomical detransposition of the great arteries was the commonest operation (N = 226). Correction was complete (biventricular) in 97% of cases (317/326). Twenty-eight neonates died. The operative mortality was 8.5%, much higher than that of children of 3 months or over (1.5%; p < 0.0001). There were 8 late deaths, all in the first postoperative year. The overall medium-term survival rate was 88.9 +/- 1.7%. It was 100% after correction of truncus arteriosus without interruption of the aortic arch; 94.2 +/- 1.5% after anatomical detransposition; 85.7 +/- 9.4% after commissurotomy of aortic stenosis; 79.2 +/- 8.3% for all the complex forms of interruption of the aortic arch; 75 +/- 9.7% for total anomalous pulmonary venous drainage and 42.9 +/- 18.7% after the Norwood palliative procedure of hypoplastic left heart syndrome. The authors conclude that early cardiac surgery saves the large majority of neonates suffering from complex congenital cardiac disease with poor prognosis. Survival stabilises one year after the operation. Other techniques or treatments are necessary to lower present surgical risk (8.5%) to that of surgery under cardiopulmonary bypass of children over 3 months of age (1.5%).

Objective: Outcome after correction of atrioventricular septal defect depends to a great deal on the postoperative function of the left atrioventricular valve. The related role of the zone of apposition ('cleft') has been debated: should it be closed (bileaflet repair) or should it be left untouched (trileaflet repair)? This study aims to answer the question by comparing the outcome of patients treated according to these two approaches. Methods: We reviewed all our patients who underwent repair of complete atrioventricular septal defect from 1984 to 1997 and selected those in whom the closure of the zone of apposition in principle would have been possible. Two groups with similar characteristics were constituted: group I (n = 63), where the zone of apposition was deliberately not closed as part of a trileaflet repair (postoperative open zone of apposition) and group II (n = 96), where it was electively closed as part of a bileaflet AV valve repair (closed zone of apposition). Since we changed from a trileaflet to a bileaflet repair in 1987, the two groups differ in terms of size and length of follow-up. Outcome was compared with regard to survival and freedom from reoperation for left atrioventricular valve incompetence. Late atrioventricular valve function was evaluated by Fcho-Doppler. For statistical analysis, we used Chi-square or Fisher's exact test, the Mann-Whitney test and the log-rank test for comparison of Kaplan-Meier curves. The difference was considered statistically significant with a P-value of 0.05 or less. Results: Early mortality was 9.5% (6/63) in group I and 3.1% (3/96) in group n (P = 0.16). Actuarial survival after 1, 4 and 8 years was 80.4, 68.4 and 64.8%, respectively, for group I. Actuarial survival for group II was 94.7, 92.1 and 92.1% (P = 0.0002). Freedom from reoperation for left atrioventricular valve regurgitation was 90.2, 85.6 and 77.8% for group I at the same time interval. It was a constant 97.9% for group II (P = 0.0016). At reoperation, left atrioventricular valve regurgitation was present through the open zone of apposition in 63% of group I cases. The follow-up is 96% (126/131) complete. An increase in degree of left atrioventricular valve incompetence was noted in 28% (11/39) of group I cases and in 9% (8/87) of group II cases (P = 0.0131). Conclusion: This study demonstrates the advantage of closing the zone of apposition ('cleft') as part of repair of complete atrioventricular septal defect. Survival, freedom from reoperation for left atrioventricular valve incompetence and over-all outcome were more favourable in patients of group n. The zone of apposition should be surgically addressed whenever the morphology of the left atrioventricular valve allows for closure without producing stenosis. (C) 2000 Elsevier Science B.V. All rights reserved.

Most publications on atrial isomerism are autopsy or case reports. The authors review 41 consecutive children operated on from 1980 through to 1996 with emphasis on associated cardiac anomalies, surgical procedures and outcome. Left atrial isomerism was present in 23 patients. Interruption of the inferior vena cava (56%). atrio-ventricular septal defect (47%), common atrium (38%) and cor triatriatum sinistrum (30%) were the most common diagnoses. Biventricular repair was achieved in 17 children and total cavo-pulmonary connection in two. Three underwent staged palliation: modified Blalock—Taussig shunt for two and bidirectional Glenn anastomosis for one. The remainder received a cardiac pacemaker. One patient died early after repair. Two underwent reoperation to correct a regurgitant left atrio-ventricular valve: one of these, in another hospital, had peroperative death. Three died later. Actuarial survival rate after repair and total cavo-pulmonary connection that was stabilized after 2 years was 84%. In the 18 children with right atrial isomerism, pulmonary atresia or stenosis predominated (89%) with discordant ventriculo-arterial connection (72%). atrio-ventricular septal defect (72%), ‘single’ ventricle (55%) and extracardiac total anomalous pulmonary venous drainage (50%). Biventricular repair was achieved in two patients and complete Fontan circulation in eight. The other eight underwent various staged palliative procedures and correction of extracardiac total anomalous pulmonary venous drainage. Five patients died postoperatively: two in our unit after modified Blalock-Taussig shunt and total cavo-pulmonary connection, three In other hospitals after repair (n = 1) and Fontan (n = 2). Five died later. One was lost for review. Survival after repair and Fontan stabilized after 6 months at 49%. In conclusion, the cardiovascular malformations associated with left atrial isomerism can often be successfully corrected. Those accompanying right atrial isomerism usually preclude a biventricular repair, require staged palliation and carry a poor prognosis.