Non-invasive cardiac output monitor validation study in pediatric cardiac surgery patients.

STUDY OBJECTIVE: To validate a non-invasive cardiac output monitor in pediatric cardiac surgery patients. DESIGN: Prospective trial. SETTING: Operating room. PATIENTS: 20 pediatric cardiac surgery patients ASA physical status 3 and 4. INTERVENTIONS: Aesculon noninvasive cardiac output monitor was used and compared to the cardiac output derived from the Fick equation. Oxygen consumption was measured and blood samples were taken from the arterial line and from the superior and inferior vena cava. MEASUREMENTS: Noninvasive cardiac output and cardiac index and Fick cardiac output and cardiac index. Oxygen consumption was measured by the TreyMed metabolic monitor. Blood samples were simultaneously drawn from the arterial line and from the superior and inferior vena cava purse string sites by the surgeon, prior to commencing cardiopulmonary bypass. Another data set was obtained right after termination of cardiopulmonary bypass. RESULTS: There was a direct, significant relationship between Fick CO/CI and NICOM CO/CI measurements. More dispersion is detected when the magnitude of the measure increases, i.e., for older and larger patients. CONCLUSIONS: There is a strong correlation between the cardiac output values derived from the Fick equation and the Aesculon non-invasive cardiac output monitor.

Evaluation of Residual Coarctation in Infants with a Single Right Ventricle after Stage I Palliation.

In infants with a single right ventricle (RV), stage I palliation involves aortic reconstruction, systemic-to-pulmonary shunt placement, and atrial septectomy. Many require re-intervention for residual coarctation of the aorta (CoA). Doppler echocardiography can detect residual CoA in most infants, but its ability to predict severity has not been studied. This study compares gradients from Doppler interrogation to those from cardiac catheterization in infants with residual CoA. We performed a retrospective study of infants after stage I palliation from 2000 to 2014. Infants with an echocardiogram and catheterization before the second-stage palliative surgery were included. Infants with an echocardiogram >30 days before catheterization were excluded. Doppler-derived gradients were compared to catheterization-derived gradients. Echocardiographic assessment of tricuspid valve (TV) and RV function were recorded. The cohort included 95 infants, and thirty-three (35%) had CoA. Doppler-derived and catheterization-derived gradients correlated weakly in infants with CoA (r = 0.37, p = 0.036) and without CoA (r = 0.35, p = 0.005). Among infants with CoA, 17/33 had none or trivial tricuspid regurgitation (TR) and normal RV function, and Doppler-derived gradients correlated with catheterization gradients in this group (r = 0.71, p = 0.001). In 16/33 infants with ≥moderate TR or RV dysfunction, gradients did not correlate (r = -0.003, p = 0.992). After a stage I palliation in infants with single RV and CoA, Doppler-derived gradients poorly predicted the severity of CoA. Infants with normal TV or RV function had Doppler-derived gradients more predictive of catheterization-derived gradients. Doppler-derived gradients have limited utility in determining the severity of CoA after a stage I palliation.

Risk Factors for Longer Hospital Stay Following the Fontan Operation.

OBJECTIVES: Children with functional single ventricle undergoing the Fontan operation consume considerable resources. The purpose of this study is to evaluate pre- and intraoperative risk factors for longer hospital stay and to describe the perioperative course at a single institution over a 15-year period. DESIGN: Retrospective cohort study. SETTING: A single pediatric cardiac ICU. PATIENTS: All consecutive patients undergoing a first-time Fontan operation from 2000 to 2014. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Prolonged length of stay was defined as hospital stay greater than 75 percentile at our institution after surgery. Of 218 patients who met inclusion criteria, median length of stay was 10 days (interquartile range, 8-14 d); prolonged length of stay was defined greater than or equal to 15 days. Independent pre- and intraoperative risk factors for prolonged length of stay included higher hemoglobin (odds ratio, 1.29; p = 0.003), higher mean pulmonary artery pressure (odds ratio, 1.14; p = 0.037), and lower aortic saturation (odds ratio, 0.92; p = 0.008) in the entire group. When patients with hepatic vein inclusion (following previous Kawashima) were excluded, higher hemoglobin (odds ratio, 1.24; p = 0.027), lower aortic saturation (odds ratio, 0.92; p = 0.017), and placement of a fenestration (odds ratio, 2.438; p = 0.021) were associated with prolonged length of stay. Fifty-eight patients (26.6%) had major complications defined by Pediatric Cardiac Critical Care Consortium. Placement of a fenestration (odds ratio, 2.297; p = 0.014) and longer aortic cross-clamp time (odds ratio, 1.015; p = 0.003) were independently associated with Pediatric Cardiac Critical Care Consortium major complications. CONCLUSIONS: In this series, 75% of patients had a postoperative length of stay less than or equal to 2 weeks. Preoperative factors suggesting worse hypoxemia/decreased pulmonary blood flow (higher hemoglobin and lower oxygen saturation) and increased pulmonary artery pressure were associated with prolonged length of stay. These findings may help risk stratify this complex patient population, provide more accurate family counseling, and provide preliminary data for changes in preoperative timing of the Fontan and/or changes to postoperative management strategies for those at high risk for increased ICU morbidity.

Echocardiographic and Surgical Correlation of Coronary Artery Patterns in Transposition of the Great Arteries.

OBJECTIVE: Determine the accuracy of echocardiography to diagnose coronary anatomy in transposition of the great arteries and to evaluate the effect of accuracy on surgical outcomes and changes in accuracy over time. DESIGN: Retrospective chart review of neonates admitted February 1999 to March 2013 with transposition. Coronary pattern from the preoperative echocardiogram and operative reports were collected and compared with determine diagnostic accuracy. Coronary patterns were further confirmed by intraoperative images taken during surgery. SETTING: Tertiary care children's hospital. PATIENTS: Neonates with transposition of the great arteries and planned arterial switch operation with an echo and operative report or image describing the coronaries. INTERVENTIONS: Not applicable. OUTCOME MEASURES: Accuracy of echocardiography to diagnose coronary anatomy in transposition, and to identify factors related to correct diagnosis. RESULTS: One hundred forty-two patients met inclusion criteria with 122 correctly diagnosed, 16 incorrect, and 4 inconclusive. Accuracy was 86%, with 95% accuracy in patients with typical coronary patterns, 85% with the most common variant (left coronary from the leftward sinus and right and circumflex from the rightward sinus), and 61% with less common patterns. Typical and common variants were more likely to be correct than atypical patterns (P < .001). Cases with ventricular septal defect were more likely to have correctly diagnosed coronaries than with an intact ventricular septum (94% vs. 79%, P = .01). There was no change in accuracy over time (P > .05). There was no difference in duration of cardiopulmonary bypass, cross-clamp times, length of stay, or postoperative stay between the correct and incorrectly diagnosed groups (P > .05). CONCLUSIONS: In our center, accuracy of echocardiographic imaging of the coronary arteries in transposition was 86% without improvement over time, and perioperative outcomes were not affected by diagnostic accuracy. Further invasive imaging may not be necessary to determine the coronary pattern in this lesion.

Pediatric Mechanical Support with an External Cardiac Compression Device.

The PediBooster external cardiac compression device is a minimally invasive, non-blood contacting Biventricular Assist Device (BiVAD) intended for pediatric use. It is being developed as a palliative therapy for acute Postcardiotomy Shock (PCS). The PediBooster extracardiac wrap is pneumatically actuated to circumferentially compress the heart, providing co-pulsation support. Attachment is via a novel hydrogel coating. Early versions of the wrap were tested in vivo using a single ventricle congenital heart disease model with postcardiotomy shock, which proved unstable and demonstrated high peri-operative mortality. The final wrap design was tested in 4 acute studies with piglets (5.1 ± 0.3 kg), where the combination of ASD and PA banding induced acute right ventricular dysfunction. Data collected included routine hemodynamic values, TEE, video of the exposed heart, and cardiac histology. The model proved stable for support durations ranging from 2 to 16 hours. The wrap restricted the heart in 3 of the 4 animals, as evidenced by increased diastolic LVP during support compared to the baseline failure condition. TEE and video data showed good attachment and function of the wrap, particularly during the final 16 hr study. This model of congenital heart disease shows promise for chronic (24-72 hr) studies. Ventricular filling during support may be improved by adjusting wrap dimensions to eliminate end diastolic restriction.

Comparison of point-of-care activated clotting time systems utilized in a single pediatric institution.

This study compares four different activated clotting time (ACT) point-of-care (POC) testing systems used at our institution for the management of patients undergoing heparin therapy. We evaluated these systems under identical conditions to determine their accuracy, reproducibility, ease of use, and cost. Two separate testing stations containing four ACT systems were used. The testing order was randomized for every sample and performed by two trained individuals. Samples of fresh heparinized whole blood were taken at regular intervals and distributed to each station. Each operator tested 50 samples, totaling 400 ACT tests. The ACT value was significantly affected by the type of machine used at both stations 1 and 2 (p < .001). Compared with all systems, the Medtronic ACT Plus Automated Coagulation Timer System (ACT Plus) resulted in the most consistent ACT values (median = 171, Interquartile Range (IQR): 169-175) and least variability (172.17 +/- 5.24). The Hemochron Signature Elite Whole Blood Microcoagulation System had the most variability (221.10 +/- 14.78) and yielded consistently higher ACT values (median = 220, IQR: 210-229.5) compared with other systems. The ACT values reported by the i-STAT Handheld and Test Cartridge Blood Analysis System (153.30 +/- 7.87) were consistently lower (median = 154, IQR: 147-161) in comparison to the ACT Plus and Medtronic HMS Plus Hemostasis Management System (180.60 +/- 7.60, median = 181, IQR: 175-186). There was no statistical difference in results between the two testing sites (p > .05) or the operators (p > .05). The significant finding of this study was the affect each system has on the ACT value. This investigation demonstrates the variability that exists among different ACT monitoring systems at our institution. The discrepant variation in ACT values that exists with the Hemochron system questions the reliability of its use in the management of patients undergoing heparin therapy.

Hybrid aortic reconstruction for treatment of recurrent aortic obstruction after stage 1 single ventricle palliation: medium term outcomes and results of redilation.

OBJECTIVE: We describe a hybrid approach to the treatment of aortic obstruction after stage 1 palliation (S1P) of hypoplastic left heart syndrome. BACKGROUND: Recurrent aortic obstruction is a common problem after S1P of hypoplastic left heart syndrome. Even mild aortic obstruction is poorly tolerated so early and definitive therapy is desirable. Although stent implantation is an effective treatment for aortic obstruction in older children and adults, technical issues due to small vessels and concerns regarding future potential for expansion have generally precluded the use of stents in this setting. METHODS: Six patients underwent hybrid aortic reconstruction (HAR) in the operating room or catheterization laboratory, with the interventional cardiologist and cardiac surgeon working in cooperation. RESULTS: Patients had a mean weight of 5.8 kg (2.9-7.7) and a mean age of 5.6 months (0.5-12.9) at the time of HAR. Five patients had undergone prior balloon angioplasty at a mean age of 2.8 months (2.1-3.5), and five had moderately depressed single ventricular function prior to HAR. The balloons used had a diameter of 7-10 mm and introducer sheath size ranged from 6 to 10 F. There were no immediate or late procedure related complications. Stent redilation was performed in 5 patients for relief of recurrent obstruction or to keep pace with somatic growth. At a median follow up of 4.8 years (0.2-7.9), there were 3 patients progressing well after Fontan palliation and 3 deaths. CONCLUSIONS: HAR allows for placement of stents that can ultimately reach adult size in small infants who have recurrent aortic obstruction after balloon angioplasty following S1P. Advantages include freedom from delivery sheath constraints when determining stent type/size, facilitation of precise stent position, and avoidance of vascular damage or hemodynamic compromise during the procedure. Longer follow-up and larger experience are required to determine if this therapy will provide a long-term solution to this difficult problem.

The Fontan operation: the pursuit of associated lesions and cumulative trauma.

Our programmatic approach to the Fontan operation has evolved to include using an extracardiac conduit with aggressive presumptive treatment of associated lesions either in the catheterization laboratory or the operating room. Fenestration is used selectively based on hemodynamics, anatomy, and presence of associated lesions. We reviewed our experience to determine the effectiveness and outcome of this strategy and to assess the cumulative trauma to the patients. The records of 137 consecutive patients who underwent Fontan at Miami Children's Hospital from 1995 to 2008 were reviewed. At mean follow up of 5.76 years, freedom from death or transplantation is 94.2% (129/137). Median age at operation was 4.6 years. Longer length of stay correlated with older operative age (P = 0.0056). Pacemakers were implanted in 11.7% (16/137). Additional (not pre-Glenn or pre-Fontan) interventional catheterizations were performed in 51.8% (71/137). Additional operations were done in 10.2% (14/137). No patient has required replacement or revision of the extracardiac conduit. Our current approach to the Fontan operation provides acceptable midterm results. The pursuit of residual lesions results in a significant number of additional interventional catheterizations and operative procedures but might have an important influence on long-term survival after the Fontan procedure.

An institutional approach to, and results for, patient with tetralogy with pulmonary atresia and major systemic-to-pulmonary collateral arteries.

BACKGROUND: Tetralogy of Fallot with pulmonary atresia and diminutive or absent intrapericardial pulmonary arteries is a rare congenital abnormality, with high morbidity and mortality. Despite great advances in surgical- and catheter-based therapies, management remains challenging and controversial. We describe the surgical methods and the results from our institution. METHODS: We performed a retrospective study of the medical records of patients included in our institutional database with tetralogy and pulmonary atresia, concentrating on those predominantly managed by our programme over their lifetime. We obtained demographics and records of all catheterisations and operations, and established mortality. We assessed the current state of those surviving in terms of clinical function at their most recent clinical evaluation and right ventricular function by echocardiography. RESULTS: We assessed 38 patients, with 89% follow-up. The mean number of catheterisations for each patients was 5, with a range from 1 to 15. The mean number of operations was 2.2, with a range from 1 to 6. Unifocalisation had been performed in 26 patients, with 12 undergoing procedures to recruit the native pulmonary vasculature. Of the overall cohort, eight patients died. The ventricular septal defect had been closed in all but two patients. Most patients have no or mild exercise intolerance. Right ventricle dysfunction has been a continuing hazard for 15 years. CONCLUSIONS: An individualised approach, using unifocalisation as well as aggressive attempts to recruit the available native pulmonary vasculature, achieves outcomes in the intermediate term superior to the natural history of the lesions, and comparable with those of other studies.

Urgent permanent pacemaker implantation in critically ill preterm infants.

The management of complete heart block in premature low birth-weight infants, particularly those with hydrops fetalis, is challenging. We report emergent implantation of permanent epicardial pacemakers in the first 48 hours of life in two premature infants (one with hydrops fetalis) with birth weights of 1,400 grams and 1,000 grams.

Transforming patient and family access to medical information: utilisation patterns of a patient-accessible electronic health record.

OBJECTIVE: The purpose of this study was to evaluate the utilisation of a web-based multimedia patient-accessible electronic health record, for patients with congenital cardiac disease. PATIENTS AND METHODS: This was a prospective analysis of patients undergoing congenital cardiac surgery at a single institution from 1 September, 2006 to 1 February, 2009. After meetings with hospital administration, physicians, nurses, and patients, we configured a subset of the cardiac program's web-based clinical electronic health record for patient and family access. The Electronic Health Record continuously measured frequency and time of logins, logins during and between hospitalisations, and page views by type (imaging versus textual data). RESULTS: Of the first 270 patients offered access to the system, 252 became users (93% adoption rate). System uptime was 99.9%, and no security breaches were reported. Users accessed the system more often while the patients were in hospital (67% of total logins) than after discharge (33% of total logins). The maximum number of logins by a family was 440, and the minimum was 1. The average number of logins per family was 25. Imaging data were viewed significantly more frequently than textual data (p 0.001). A total of 12 patients died during the study period and 11 members of their families continued to access their Electronic Health Records after the date of death. CONCLUSIONS: A web-based Patient Accessible Electronic Health Record was designed for patients with congenital cardiac disease. The adoption rate was high, and utilisation patterns suggest that the Electronic Health Record could become a useful tool for health information exchange.

Continuous incisional infusion of local anesthetic in pediatric patients following open heart surgery.

AIM: To determine the efficacy and safety of a continuous subcutaneous local anesthetic (LA) infusion in pediatric patients following open heart surgery. BACKGROUND: The use of a continuous LA infusion has been shown to be beneficial following adult cardiac surgery. To date there are no studies in the pediatric population. METHODS/MATERIALS: Using a prospective, randomized, and double blind design, we compared LA, either 0.25% levobupivacaine or bupivacaine (Treatment Group) to saline (Placebo Group) delivered subcutaneously via a continuous infusion for 72 h after open heart surgery in 72 patients. Requirements for postoperative analgesics and pain scores were recorded for 72 h and plasma levels of local anesthetic were measured. Secondary outcomes measures included time to first oral intake, time to first bowel movement, time to urinary catheter removal, length of stay, requirements for antiemetics and additional sedation. RESULTS: Total morphine requirements over the first 24 h were less in the Treatment Group than the Placebo Group (0.05 mg x kg(-1) vs 0.2 mg x kg(-1), P = 0.007); this was true for all patient groups except those patients weighing less than 6.3 kg. The number of patients requiring no morphine was greater in the Treatment Group (7/35 vs 1/37, P = 0.02). The Treatment Group also received less midazolam, lorazepam, and ketorolac than the Placebo Group over 72 h due to the reduced clinical need for these agents in patients weighing less than 31 kg. There were no differences in secondary outcomes. CONCLUSIONS: A continuous incisional infusion of LA reduced postoperative analgesic requirement and sedative use in pediatric patients undergoing a median sternotomy incision. Dosed at a maximum rate of 0.4 mg x kg(-1) x h(-1), a continuous incisional infusion of LA is effective and safe for up to 72 h, with plasma levels of local anesthetic well below the toxic threshold.

Hybrid ventricular decompression in pulmonary atresia with intact septum.

Initial palliation for pulmonary atresia with intact ventricular septum continues to evolve in the face of significant early and late morbidity. In patients with suitable anatomy, decompression of the right ventricle may be the first step in treatment. A hybrid approach to right ventricular decompression, combining surgery and interventional catheterization techniques is described. Direct access to the right ventricle through a subxiphoid incision with transventricular sheath placement is used to provide optimum catheter position for radiofrequency perforation of membranous pulmonary atresia followed by balloon dilation. The technique may address key limitations of the traditional surgical and interventional approaches.

Amiodarone for pediatric resuscitation: a word of caution.

Intravenous administration of amiodarone has recently been recommended for use during pediatric resuscitation of pediatric patients with ventricular fibrillation or pulseless ventricular tachycardia. We present two pediatric patients who received amiodarone for polymorphic ventricular tachycardia, although they were ultimately determined to have congenital long QT syndrome. Amiodarone is contraindicated in this setting and may have exacerbated the ventricular arrhythmia.

Hybrid treatment of superior vena cava syndrome in a child.

A 10-year-old boy with a history of renal failure and hemodialysis by indwelling superior vena cava (SVC) catheters was diagnosed with SVC obstruction and clinically severe SVC syndrome. During attempted recanalization of the SVC in the cardiac catheterization laboratory, he suffered a perforation of his SVC with pericardial tamponade. After treatment of the perforation and relief of tamponade, he underwent a hybrid procedure to recanalize his SVC. A needle and then guidewire were passed directly from the right atrium through the SVC obstruction and were used to successfully dilate and stent the obstruction.

Midterm results for collaborative treatment of pulmonary atresia with intact ventricular septum.

BACKGROUND: We report a single-institution experience using a collaborative surgical and catheter-based approach to the initial treatment of pulmonary atresia with intact ventricular septum. METHODS: A retrospective review was conducted of all neonates admitted with pulmonary atresia with intact ventricular septum from 1996 to March 2007. RESULTS: We identified 24 patients with a mean age at first intervention of 4.5 days with mean follow-up of 6.05 years (range, 1.9 to 12.7 years). Initial palliation was determined by right ventricular size, morphology, and presence or absence of right ventricular-dependent coronary circulation. Initial catheter-based pulmonary valve perforation and valvuloplasty was performed in 41.7% (10 of 24 patients; group A), and 58.3% (14 of 24 patients) had an initial systemic-to-pulmonary artery shunt (group B). Tricuspid valve size was significantly smaller in group B (median z-score, -0.52 group A versus -2.40 group B; p < 0.001). Placement of a shunt after valvuloplasty in group A was required in 70.0% (7 of 10 patients). There was no mortality in group A, and 70.0% (7 of 10 patients) are in a two-ventricle pathway and 30.0% (3 of 10 patients) are in a 1.5-ventricle pathway. Group B had mortality of 14.3% (2 of 14 patients), both within 5 days of surgery. All group B patients remain in a single-ventricle pathway. Overall survival is 91.7% (22 of 24 patients). CONCLUSIONS: An individualized approach to this complex lesion has good results. If the right ventricle can be safely decompressed and appears usable, the need for a shunt after valvuloplasty does not preclude two-ventricle (or 1.5-ventricle) repair. Anatomy mandating a shunt as initial palliation has substantial early mortality.

Frequency and indications for tracheostomy and gastrostomy after congenital heart surgery.

Patients undergoing congenital heart surgery may occasionally require additional surgical procedures in the form of tracheostomy and gastrostomy. These procedures are often performed in an attempt to diminish hospital morbidity and length of stay. We reviewed the Web-based medical records of all patients undergoing congenital heart surgery at Miami Children's Hospital from February 2002 through August 2007. Patients who were deemed preterm and had undergone closure of a patent ductus arteriosis were eliminated. The records of all other patients were queried for the terms gastrostomy, g-tube, Nissan, fundal plication, tracheostomy, or tracheotomy. Patients' medical records in which these terms appeared in any portion were completely reviewed. There were 1660 congenital heart operations performed in the study period. There were 592 operations performed on patients whose age ranged from 1 month to 1 year and 441 neonatal operations. Mortality was 2%. Median postoperative stay was 8 days (range, 1-191 days), 12 days for neonates (range, 3-142 days), and 19 days for neonates undergoing RACHS-1 category 6 operations (range, 4-142 days). Tracheostomies were performed in four patients (0.2%). Gastrostomies were performed on eight patients (0.4%), representing 0.8% of patients <1 year of age, 1.4% of neonates, and 2.4% of patients undergoing RACHS-1 category 6 operations. The rate of patients undergoing either tracheostomy or gastrostomy after congenital heart surgery at our institution was quite low. Avoidance of either of these two procedures was achieved without increased morbidity or length of stay. The rate at which these procedures need to be performed may reflect the magnitude of the patients' lifetime trauma related to their underlying condition and acute and total surgical experiences.

Congenital cardiac surgery without routine placement of wires for temporary pacing.

OBJECTIVE: Temporary pacing wires have been associated with serious postoperative complications. Recommendations for their routine use after open heart surgery are decades old, and may not reflect current surgical techniques and outcomes. METHODS: The electronic web-enabled medical records of all patients undergoing congenital cardiac surgery from February, 2002, through December, 2005, were reviewed, excluding patients undergoing implantation of pacemakers as a primary procedure, or those undergoing ligation of a patent arterial duct. RESULTS: There were 1193 surgical procedures performed, 1041 with cardiopulmonary bypass. Median age of the patients was 5.8 months, with a range from 0 days to 54 years, weighing 6.2 kilograms, with a range from 1 to 114 kilograms. Mortality prior to discharge was 2.5%, and median postoperative stay was 6 days. No deaths were attributed to arrhythmias. Temporary pacing wires were placed 14 times (1.2%). Indications for placement included sinus nodal dysfunction in 8 patients, preoperative in 4 and intraoperative in 4, high degree atrioventricular block in 4 patients, and intraoperative atrial flutter in 2 patients. Of these patients, 4 (0.3%) eventually underwent permanent implantation of a pacemaker, 2 for persistent sinus nodal dysfunction, and 2 for persistent atrioventricular block. Postoperative junctional ectopic tachycardia requiring antiarrhythmic therapy occurred in 9 patients (0.8%). All recovered without incident, and none were treated with temporary pacing. CONCLUSIONS: The diminished risk of unexpected postoperative arrhythmias in the current era alleviates the necessity for routine placement of temporary pacing wires. Those institutions with experienced surgical and cardiac critical care teams may be able to predict the need for temporary pacing wires preoperatively or intraoperatively.

Rapid cardiopulmonary support in children with heart disease: a nine-year experience.

BACKGROUND: We developed a novel mechanical rapid cardiopulmonary support system (CPS) in 1996 to eliminate what we believed were shortcomings of conventional extracorporeal membrane oxygenation (ECMO) circuits when used in patients with congenital heart disease. We reviewed the use of this system over a nine year period to determine if we had been successful in improving results compared with ECMO and if outcomes have changed over this time. METHODS: All children supported with CPS (110 procedures) were reviewed. Noncardiac CPS cases (7) were excluded. The study population was divided into two time periods (1995 to 2000 and 2001 to 2004), which correlate with significant differences in intraoperative, postoperative, and CPS management. Patients were further analyzed by age (< or = 30 days or > 30 days), repair complexity (risk adjusted classification for congenital heart surgery [RACHS]-1 category 6 or categories 1 to 5), and length of support. RESULTS: Overall thirty day survival of cardiac CPS patients was 55% (57 of 103). Overall survival increased from 45% (23 of 51) during the first period to 65% (34 of 52) during the second period [p < or = 0.005]. Survival rates in neonates improved from 41% (11 of 27) to 56% (15 of 27) and RACHS-1 category 6 survival improved from 38% (5 of 13) to 69% (9 of 13), but neither change reached statistical significance. Intracranial hemorrhage occurred in 6.4% of all CPS patients. CONCLUSIONS: Cardiopulmonary support is an effective alternative to ECMO for pediatric cardiac support. Further, our experience suggests that patient survival may be improved by CPS compared with reported results for ECMO in cardiac patients.