Preoperative and postoperative evaluation of airways compression in pediatric patients with 3-dimensional multislice computed tomographic scanning: effect on surgical management.

OBJECTIVES: Surgical management of airway compression of vascular origin requires an accurate analysis of anatomy and various mechanisms of compression. This study assessed the usefulness of 3-dimensional computed tomographic scanning in the preoperative and postoperative evaluation of airways compression in a pediatric population. METHODS: Thirty-seven consecutive patients (median age, 4 months) were examined with multislice 3-dimensional computed tomographic scanning: 18 patients before surgical treatment of anomalies of vascular rings, 2 patients because of respiratory symptoms after repair of esophageal atresia, and 17 patients because of persisting respiratory symptoms or prolonged mechanical ventilation after cardiac surgery for congenital heart disease. RESULTS: The procedure was successful, with high-quality diagnostic imaging obtained in all cases without any complications. The anatomy and relationship between the vascular arches and airways was analyzed in all referred patients with vascular arch anomalies confirmed on the basis of the surgical findings, and this helped the surgeon to plan the procedure and choose the best approach. After cardiac surgery, the airway and vascular structures involved and the mechanism of compression were specified in all but one case, and the 3-dimensional computed tomographic scan serves as an important tool for deciding whether to perform reoperation on patients requiring prolonged mechanical ventilation. CONCLUSION: Three-dimensional computed tomographic scanning is a safe, fast, and noninvasive method useful for accurately analyzing the mechanisms of airway compression of vascular origin and thus possible improving the surgical management of pediatric patients.

Surgery for right ventricle to pulmonary artery conduit obstruction: risk factors for further reoperation.

OBJECTIVE: To identify the surgical approaches and risk factors which influence longevity of right ventricle to pulmonary artery (RV-PA) conduits following first reoperation for obstruction. METHODS: Between January 1993 and August 2003, 114 patients underwent 141 reoperations for RV-PA conduit obstruction. Diagnoses included 'Truncus Arteriosus' (n=52), 'Pulmonary atresia/Tetralogy of fallot' (n=39), 'Double outlet right ventricle' (n=10), 'Transposition of great arteries, VSD, and pulmonary atresia' (n=9), and the 'Ross operation' (n=4). All patients had undergone a previous biventricular repair. The first reoperation for conduit obstruction was performed in 112 hospital survivors by: total conduit replacement (Group A, n=73) with valved (homograft=10 and xenograft=54) or non-valved (n=9) conduit, and patch enlargement of the obstructed RV outflow tract with preservation of the posterior and sides of the conduit wall after removing of the fibrocalcific peel and degenerated valve (Group B, n=39). Mean age at first reoperation was 8.8+/-6.7 and 7.5+/-5.3 years in patients of groups A and B, respectively. Seven patients in Group A and 18 in Group B required a second reoperation and two patients in Group B a third reoperation. RESULTS: There were two hospital deaths and no late deaths. Mean follow-up was 5.8+/-3.2 years. Risk factors for second reoperation by univariate analysis were: homograft conduit use (P=0.004), Group B surgical approach (P=0.0001), higher RV-PA systolic pressure gradient at discharge (P=0.02), and age <5-years-old (P=0.01). Multivariate analysis showed that inclusion in Group B and younger age (<5-years-old) at repair were independent risk factors for second reoperation. Group B surgical approaches had higher RV-PA systolic pressure gradient at discharge (P=0.02) and required more PA bifurcation repair at the time of second reoperation (P=0.05). Freedom from second reoperation for conduit obstruction was significantly higher in Group A patients at 5 and 8 years (P<0.04) and those with xenografts rather than homograft (P=0.04). CONCLUSIONS: Our results support the optimal surgical approach for RV-PA conduit obstruction is total replacement with a xenograft. RV outflow reconstruction by other techniques without complete dissection of PA bifurcation does not completely relieve the stenosis and could cause early restenosis. Higher systolic gradients at discharge and younger age at first reoperation are predictors of earlier reoperation.

Pulmonary vascular endothelial growth factor and nitric oxide interaction during total cardiopulmonary bypass in neonatal pigs.

OBJECTIVE: Lung injury after cardiopulmonary bypass includes pulmonary hypertension and lung edema. Both complications are related to endothelial pulmonary vascular dysfunction, leukocyte sequestration, and increased capillary permeability. This study was done in an attempt to better define the endothelial dysfunction and the cause of edema. METHODS: Twenty-five neonatal piglets were subjected to total cardiopulmonary bypass for 90 minutes without crossclamping of the aorta. After weaning from cardiopulmonary bypass, they were allowed to survive 2 hours, at which time they were killed. Preoperative and postoperative hemodynamic studies, lung (n = 16) and muscular (n = 5) vascular endothelial growth factor contents, and exhaled nitric oxide (n = 8) were recorded. Immediately after the animals were killed, pulmonary arterial rings were obtained from 12 piglets and mounted in organ chamber for assessment of endothelial function with receptor-dependent (acetylcholine) or non-receptor-dependent (calcium ionophore A23187) studies and compared with control pulmonary arterial rings. The left lungs of 13 piglets were mounted in isolated perfused lung chambers for filtration coefficient assessment and comparison with control preparations. RESULTS: After cardiopulmonary bypass, pulmonary vascular resistance increased from 953.7 +/- 302.6 dyne x s x cm(-5) to 1973.6 +/- 925.4 dyne x s x cm(-5) (P =.03). This was associated with an increase in lung vascular endothelial growth factor content from 91.07 +/- 5.314 pg/100 mg tissue to 151.6 +/- 11.4 pg/100 mg tissue (P <.0001), an increase in muscle vascular endothelial growth factor from 76.02 +/- 11.53 pg/100 mg tissue to 81.58 +/- 7.7 pg/100 mg tissue (P not significant), and a decrease in exhaled nitric oxide from 6 +/- 1.7 ppb to 3.12 +/- 1.4 ppb (P =.003). The filtration coefficient was statistically significantly higher after cardiopulmonary bypass than in control preparations (0.259 +/- 0.02 vs 0.525 +/- 0.07, P <.0001). Variations in lung vascular endothelial growth factor accumulation were statistically significantly higher than in muscular vascular endothelial growth factor accumulation (60.5 +/- 9.1 vs 5.5 +/- 5.9, P =.0008). In addition, a statistically significant correlation was found between postbypass lung vascular endothelial growth factor and lung filtration coefficient (P =.0058), as well as between change in lung vascular endothelial growth factor and change in lung filtration coefficient (P =.03). Pulmonary vascular endothelial receptor-dependent (acetylcholine) function was statistically significantly blunted after bypass relative to control values (15.44% +/- 4.8% vs 55.5% +/- 5.96% of maximal relaxation, P =.0001), whereas non-receptor-dependent endothelial function was unaffected by cardiopulmonary bypass (110.77% +/- 8.9% vs 120.63% +/- 15.46% of maximal relaxation, P not significant). CONCLUSIONS: These findings suggest that lung ischemia that occurs during cardiopulmonary bypass affects the signal transduction from membrane receptors to intracellular calcium mobilization and nitric oxide synthase activation. Lung edema after bypass is probably due in part to lung accumulation of vascular endothelial growth factor, a finding that was not found in systemic muscular nonischemic territories.

Is surgical intervention still indicated in recurrent aortic arch obstruction?

BACKGROUND: Introduction of balloon dilatation has become the standard treatment for recurrent aortic arch obstruction and has changed the therapeutic approach to patients with this disorder. OBJECTIVES: Whether all patients with recurrent aortic arch obstruction are candidates for balloon dilatation remains unanswered. In addition, only few reports have tried to compare the results between patients undergoing balloon dilatation or redo operations. METHODS: Since 1983, 97 patients underwent reintervention for recurrent aortic arch obstruction (42 dilations and 55 reoperations). Eight had immediate unsuccessful dilatation and were shifted to the surgical group (n = 63). The median age at reintervention was 21.7 months (10 days-45 years), and the median delay was 13.6 months (7 days-17 years). Anatomy of the aortic arch oriented the surgical approach to treat arch hypoplasia. It could be performed through a left thoracotomy in 52 patients, with extended end-to-end anastomosis in 34 patients, subclavian flap repair in 9 patients, conduit insertion in 6 patients, and patch enlargement in 3 patients. More recently, an anterior approach with cardiopulmonary bypass without circulatory arrest was applied to enlarge the patch in all the aortic arches. RESULTS: There was one early death in the surgical intervention group and 2 late deaths in the dilation group. Major complications and recurrence were higher in the dilated group (4 vs 0, P <.01, and 14 vs 5, P <.0004, respectively). At a mean follow-up of 11.8 +/- 4.1 years in the surgical intervention group and 7.5 +/- 2.5 years in the dilated group, systemic hypertension was normalized in all but 5 patients in the surgical intervention group and 6 patients in the dilated group. CONCLUSION: Reoperation for recurrent aortic arch obstruction can be performed safely, with low rates of mortality and morbidity. This approach should be considered versus balloon angioplasty, especially in patients older than 4 years and in the presence of aortic arch hypoplasia.

Left-sided lesions after anatomic repair of transposition of the great arteries, ventricular septal defect, and coarctation: surgical factors.

OBJECTIVE: This study was undertaken to identify potential anatomic and surgical factors creating left-sided lesions, namely recoarctation of the aorta and neoaortic regurgitation, after anatomic repair of transposition of the great arteries with ventricular septal defect and aortic coarctation. METHODS: From 1983 to September 2002, 109 survivors out of 120 patients were studied. Two-stage repair was performed in 42 patients (group A), and single-stage repair was performed in 67 (groups B and C). Before repair, the diameters of the ascending aorta and main pulmonary artery were measured. In the patients with single-stage repair, coarctation was repaired by extended end-to-end anastomosis in 35 patients (group B) and by pulmonary homograft patch augmentation in 32 patients (group C). The ventricular septal defect was closed through the pulmonary artery in 70 patients and through the right ventricle or atrium in 39 patients. The neoaorto-aortic discrepancy was treated by V-shaped resection of the posterior sinus of Valsalva in 7 cases, pulmonary homograft patch in 32 cases, and anterior splitting of the ascending aorta in all cases. Before discharge from the hospital, neoaortic root and ascending aorta diameters and aortic regurgitation grade were recorded. Neoaortic regurgitation progression and reintervention were the end points of follow-up (97.2 +/- 61.2 months). RESULTS: Early and late survivals were significantly better in group C (P <.001). Risk factors for neoaortic regurgitation at discharge by univariate analysis were single-stage repair (P <.05) and ventricular septal defect closure through the pulmonary artery (P =.0076). On multivariate analysis, the latter was the only risk factor for neoaortic regurgitation at discharge and at last follow-up. Multivariate analysis showed that higher neoaortic root/ascending aorta ratio and ventricular septal defect closure through the pulmonary artery were risk factors for neoaortic regurgitation evolution at last follow-up. There were 29 reinterventions, 19 for recoarctation of the aorta and 10 for neoaortic regurgitation with or without aortic root dilatation. Group B (P <.05), high neoaortic root/ascending aorta ratio (P <.01), and progressive neoaortic regurgitation (P <.05) were risk factors for recoarctation of the aorta. Group A was a risk factor for aortic valve replacement at 10 years (P <.05). CONCLUSION: Neonatal single-stage repair with pulmonary homograft aortic augmentation remains the optimal approach to transposition of the great arteries with ventricular septal defect and aortic coarctation. It provides better early and late survivals and freedoms from left-sided lesions. Avoidance of late recoarctation of the aorta and progressive neoaortic regurgitation requires meticulous closure of the ventricular septal defect and evenly sized reconstruction of the aorta from root to distal arch.

Aortic arch reconstruction with pulmonary autograft patch aortoplasty.

OBJECTIVE: The optimal technique for aortic arch reconstruction through median sternotomy is still under debate. We have introduced the technique of pulmonary autograft patch aortoplasty as a reliable alternative. METHODS: The outcomes of 51 infants who underwent neonatal repair of interrupted aortic arch (n = 28) or coarctation associated with ventricular septal defect (n = 23) since 1992 were analyzed. The patients were reviewed in three groups according to the aortic arch reconstruction technique: group I underwent direct anastomosis (n = 23), group II underwent homograft or pericardial patch aortoplasty (n = 8), and group III underwent pulmonary autograft patch aortoplasty (n = 20). The pulmonary autograft patch consisted in the anterior wall of the main pulmonary artery, between the supracommissural level and the divided ductus arteriosus. The created defect was replaced with fresh autologous pericardium. RESULTS: All patients except 1 were discharged without significant residual gradient at the level of the aortic arch. At a median delay of 7 months (range 2-51 months), 11 patients (22%) had recurrence of arch obstruction and underwent balloon angioplasty (n = 8) or surgical correction (n = 3). One patient who had undergone direct anastomosis required reoperation for bronchial compression. At a median follow-up of 29 months, the actuarial freedoms from recurrent arch obstruction were 81% for direct anastomosis, 28% for homograft or pericardial patch aortoplasty, and 100% for pulmonary autograft aortoplasty (P =.03 for group III vs group I and P <.0001 for group III vs group II). CONCLUSIONS: The aortic arch repair associated with pulmonary autograft patch augmentation resulted in superior midterm outcomes and therefore constitutes a reliable alternative to the direct anastomosis technique. It allowed complete relief of anatomic afterload and diminished the anastomotic tension, thus reducing the risk of restenosis and tracheobronchial compression. We observed a significantly higher rate of recurrence after patch aortoplasty with other materials.

Endocardial location of familial myofibromatosis revealed by cerebral embolization: cardiac counterpart of the frequent intravascular growth of the disease?

Myofibromatosis is a rare infantile benign neoplasia, which may involve the heart in the rare and usually fatal generalized form of the disease. Diagnosis of endocardial myofibromas was made on two surgically excised lesions of the mitral valve that were reveled by a cerebral embolization in a 12-month-old female infant. Surprisingly, the patient had no other obvious lesion of myofibromatosis. However, her father had a histologically proven neonatal history of myofibromatosis. This case confirms the likely autosomal dominant mode of inheritance of myofibromatosis. It highlights the embolization risk of the previously unreported endocardial location. We suggest that these clinically isolated non-invasive endocardial myofibromas did not represent a true visceral form of myofibromatosis. They were, rather, similar to the frequent intravascular growth of the disease.

Intermediate follow-up of a composite stentless porcine valved conduit of bovine pericardium in the pulmonary circulation.

BACKGROUND: In the pediatric population, glutaraldehyde-preserved bovine pericardium conduit containing a stentless porcine valve has been proposed as an alternative to homografts for right ventricular outflow tract reconstruction. METHODS: Between June 1996 and March 2000, a total of 55 patients, 20 with truncus arteriosus, 21 with pulmonary atresia with ventricular septal defect, and 14 with miscellaneous defects, received this conduit. Median age at implantation was 3.4 months (range, 3 days to 19 years), and 27 patients (50%) were less than 3 months old. Clinical outcome, echocardiographic data, and pathologic analysis were recorded. End points for conduit failure were conduit replacement or dilation. A mean follow-up of 27 months (range, 2 to 46 months) was available for 47 survivors. RESULTS: Procedure for conduit obstruction was required in 13 patients. The most common procedure was operation, and all but 3 patients had an unsuccessful balloon angioplasty before reoperation. Actuarial freedom from conduit dilation or reoperation was 93.6% (95% confidence limits, 82% to 99%), 81.9% (95% confidence limits, 64% to 91%), 77.8% (95% confidence limits, 39% to 78%), and 64.3% (95% confidence limits, 26% to 73%) at 1, 2, 3, and 4 postoperative years, respectively. Univariate analysis identified small conduit size as a risk factor for conduit obstruction. CONCLUSIONS: Although this new conduit was not free from progressive obstruction, our clinical results (easy to work and good valvular function) and the availability in small sizes encouraged us to use it as an alternative to small-size homografts when those were not available.

Induction of pulmonary angiogenesis by adenoviral-mediated gene transfer of vascular endothelial growth factor.

BACKGROUND: We hypothesized that gene transfer of vascular endothelial growth factor (VEGF) mediated by an adenovirus vector might induce pulmonary artery angiogenesis in a lamb model of pulmonary artery hypoplasia. METHODS: Thirteen fetal lambs had left pulmonary artery banding at 106 days of gestation. Following birth, 3 groups were divided: VEGF group (n = 5) and beta-GAL group (n = 4) received an adenoviral vector encoding respectively for human VEGF165 and for galactosidase A. A control group (n = 4) had neither gene nor virus. Viral suspensions were selectively instilled in the left bronchus 6.5 days after birth. Five nonoperated lambs constituted the normal group. Euthanasia was performed at 30 days of age. Gene transfer was confirmed by blue coloration of left lung obtained with Xgal solution in an additional experiment. Histomorphometric evaluation was performed. All groups were compared with ANOVA test and paired test was used to compare right and left lung in each animal. RESULTS: Left lung was similarly hypoplastic in all operated lambs. Left pulmonary artery hypoplasia present in all operated groups was significantly less pronounced in VEGF group. The number of pleural arteries was similarly increased in left lung of all operated lambs. Left lung arterial density was higher in VEGF group than in all other groups. The percentage of parenchyma of left lung was lower in beta-GAL group than in all others, partially returned to normal in VEGF group. CONCLUSIONS: In this model, transbronchial VEGF gene transfer induces pulmonary angiogenesis, proximal pulmonary artery growth and contributes to lung parenchyma recovery.

Biventricular repair for double-outlet right ventricle.

Double-outlet right ventricle (DORV) is a heart malformation that describes an anomalous ventriculoarterial connection which can be associated with considerable variant of associated lesions. When this malformation is present with two ventricles, biventricular repair is feasible in the vast majority of cases. This report describes the surgical techniques for biventricular repair in all forms of encountered DORV, as well as the surgical strategy employed at our institution. Copyright 2000 by W.B. Saunders Company

Results of the double switch operation for congenitally corrected transposition of the great arteries.

BACKGROUND: Congenitally corrected TGA (CC-TGA) is characterized by discordant atrioventricular and ventriculo arterial connections. In absence of right ventricular outflow tract obstruction (RVOTO), repair by atrial and arterial switches remains a challenging procedure for which long term follow-up is uncertain. METHODS: From 1995 to 2007, 20 patients (median age: 26 months) with CC-TGA had double switch procedure. Segmental anatomy was {SLL} in all patients, dextrocardia in two patients, mesocardia in two patients. Ventricular septal defect was present in 17 patients, aortic coarctation in 2 patients and interrupted aortic arch (IAoA) in 1 patient. Five patients had tricuspid valve regurgitation. Six patients had AV blocks, 4 patients had pacemaker implantation prior to repair. Pulmonary artery banding was performed in 17 patients, for congestive heart failure (14 patients) or left ventricular retraining (3 patients). Three patients, including one patient with IAoA had primary repair. After LV retraining, repair was performed when indexed LV mass to LV volume ratio was above 1.5. A median follow-up of 60 months was achieved in all. RESULTS: There were no deaths. Postoperative pacemaker implantation was required in four patients. Reoperation for Senning obstruction was necessary in one patient, and pacemaker battery replacement in another patient. One patient had mild neoaortic insufficiency, two had mild tricuspid regurgitation and two had mild mitral regurgitation. All were in NYHA I-II. Actuarial survival at 10 years was 100% and freedom from reoperation at 5 and 10 years were 93% and 77.4%, respectively. CONCLUSION: Double switch for CC-TGA without RVOTO can be performed with no mortality and low morbidity. Since these results seem to last for several years, it should be considered as the optimal procedure.

Fate of the truncal valve in truncus arteriosus.

BACKGROUND: The fate of the truncal valve (TV) after truncus arteriosus repair remains poorly defined. The purpose of this report was to analyze how the TV influences outcome of truncus arteriosus repair. METHODS: From January 1986 to December 2003, 153 patients underwent complete repair of the truncus arteriosus. Median age was 35 days. Preoperative TV insufficiency was absent or trivial in 59 patients (39%), mild in 72 patients (47%), moderate in 13 patients (8%), and severe in 9 patients (6%). The TV was quadricuspid in 36 patients (24 %), bicuspid in 10 patients (7%), and tricuspid in the remaining patients. At surgery, TV plasty (n = 6) or replacement (n = 3) was associated with truncus arteriosus repair. Truncal valve-aorta continuity was restored by patch interposition in 27 patients (17%), by conduit in 7 patients (5%), and directly in the remaining patients. RESULTS: A 97-month mean follow-up was achieved in all survivors. The actuarial survival rates were 81.7% +/- 3.1% and 79.1% +/- 3.3% at 6 months and 18 years, respectively. Among 85 patients who underwent 113 reoperations, 19 underwent 24 TV reoperations: 5 isolated TV reinterventions and 19 associated with right ventricular-pulmonary artery conduit replacement. Freedom from TV reoperation was 96%, 82.3%, and 62.7% at 1, 10, and 18 years, respectively. Multivariate logistic regression showed that moderate or severe TV insufficiency at initial presentation was a risk factor for late TV reoperation but not for early mortality. Initial TV replacement was associated with lower survival. CONCLUSIONS: Initial TV insufficiency is associated with higher reoperation rate. This condition requires more refined techniques of TV plasty. The overall long-term freedom of TV reintervention rate remains within reasonable ranges. Truncal valve reintervention presents a low risk for mortality.

Surgery for transposition of the great arteries in neonates weighing less than 2,000 grams: a consecutive series of 25 patients.

BACKGROUND: The arterial switch operation (ASO) is the optimal option for neonates with transposition of the great arteries (TGA). Low birth weight (LBW) and prematurity were considered as risk factors for poor outcome in early corrective surgery for cardiac defects. This retrospective study was undertaken to evaluate early and midterm results in infants with TGA weighing less than 2,000 grams who underwent surgical procedure in the neonatal period. METHODS: Among the 1,505 patients who underwent surgical procedure for TGA at our institution, 25 (0.02%) had a birth weight less than 2,000 grams and constituted the study group. Median age at operation was 19 days and median weight was 1,930 grams. Prior to surgery, all were in the intensive care unit. Eleven (48%) with TGA and intact interventricular septum had an ASO but one had a Senning operation. Among 13 patients (52%) with complex TGA, 9 had anatomic repair and 4 had palliation. RESULTS: Operative mortality was 16%. Age at operation greater than 30 days and palliation were risks factors for early death. At postoperative 43 months, actuarial survival rate was 71% and freedom from reoperation rate was 73%. All survivors were considered to have good cardiac status; 95% joined the normal curve for LBW infants without heart defects. CONCLUSIONS: These data support that delaying repair in LBW neonates with simple or complex TGA does not confer any benefit and is associated with higher morbidity.

Aortic valve regurgitation after arterial switch operation for transposition of the great arteries: incidence, risk factors, and outcome.

OBJECTIVES: The aims of this study were to assess the prevalence and incidence of aortic valve regurgitation (AR) after arterial switch operation (ASO), its outcome, and the risk factors. BACKGROUND: After an ASO, the long-term fate of the aortic valve is a concern as follow-up lengthens. METHODS: Operative and follow-up data on 1,156 hospital survivors after ASOs between 1982 and December 2000 were reviewed. RESULTS: At last follow-up (mean duration 76.2 +/- 60.5 months), 172 patients (14.9%) had an AR. Complex transposition of the great arteries, prior pulmonary banding done in 75 patients (21 with intact ventricular septum), aortic arch anomalies, AR at discharge, older age at ASO, and aortic/pulmonary size discrepancy were associated with AR. On multivariate analysis, the presence of a ventricular septal defect (VSD) or AR at discharge multiplied the risk by 2 and 4, respectively. Freedom from AR was 77.9% and 69.5% at 10 and 15 years, respectively; hazard function for AR declined rapidly and slowly increased thereafter. Reoperation from AR was done in 16 patients with one death, valvuloplasty being unsuccessful. Freedom from reoperation for AR was 97.7% and 96.8% at 10 and 15 years, respectively; hazard function slowly increased from 2 to 16 years. Higher late mortality was not associated with AR. CONCLUSIONS: After ASO, AR was observed and was related to VSD with attending high pressure and flow and AR at discharge. Progression of AR was slow, but incidence increased with follow-up. Reoperation for AR was rare. Late aortic valve function warrants long-term monitoring.

Additional pulmonary blood flow has no adverse effect on outcome after bidirectional cavopulmonary anastomosis.

BACKGROUND: Controversy continues over whether additional sources of pulmonary blood flow are beneficial in combination with a bidirectional cavopulmonary anastomosis. We have therefore assessed the effects of additional pulmonary blood flow on outcome after bidirectional cavopulmonary anastomosis. METHODS: From 1996 to 2000, 106 patients underwent bidirectional cavopulmonary anastomosis, either isolated (group 1, n = 54), or with additional pulmonary blood flow through the pulmonary artery (group 2, n = 30) or a Blalock-Taussig shunt (group 3, n = 22). RESULTS: Superior vena cava syndrome was more frequent in group 2 and less in groups 1 and 3 (p < 0.05). Low-output syndrome was more frequent in group 2 and less in group 3 (p = 0.01). Repeated-measures analysis of variance showed higher oxygen saturations with additional pulmonary blood flow (p < 0.05) and significant changes over time (p < 0.0001). Pulmonary pressures, systemic ventricular fractional shortening, end-diastolic diameter index, end-diastolic pressure, and atrioventricular valve regurgitation remained unaffected by additional pulmonary blood flow. Pulmonary artery pressures were lower in group 2 than 3 (p < 0.05). Fractional shortening (p < 0.05) and atrioventricular valve regurgitation (p < 0.0001) changed significantly over time. Fractional shortening showed a strong trend toward different changing patterns with or without additional pulmonary blood flow (p = 0.055), and atrioventricular valve regurgitation showed different changing patterns among the groups (p < 0.005). End-diastolic diameter and pulmonary artery dimensions, which were smaller than normal, remained unchanged. In logistic regression, smaller body surface area at bidirectional cavopulmonary anastomosis, single ventricle, and bidirectional cavopulmonary anastomosis with a Blalock-Taussig shunt were associated with early death. Actuarial survival including total cavopulmonary connection did not differ among groups (p = 0.96). CONCLUSIONS: We conclude that additional pulmonary blood flow has no adverse effect on outcome after cavopulmonary anastomosis. Additional flow through the main pulmonary artery offers different advantages and disadvantages concerning perioperative complications and pulmonary artery growth compared with additional flow through a Blalock-Taussig shunt.