HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Constantine Mavroudis Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mavroudis, C. Articles by Schwartz, M. Search for Related Content PubMed PubMed Citation Articles by Mavroudis, C. Articles by Schwartz, M.

Ann Thorac Surg 1999;68:601-624
© 1999 The Society of Thoracic Surgeons

---

Report

The Society of Thoracic Surgeons national congenital heart surgery database report:

Analysis of the first harvest (1994–1997)

^a Children’s Memorial Hospital, Chicago, Illinois, USA

Address reprint requests to Dr Mavroudis, Children’s Memorial Hospital, 2300 Children’s Plaza, MC22, Chicago, IL 60614;
e-mail: c-mavroudis{at}nwu.edu

	Abstract

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
This analysis summarizes the first report of the Society of Thoracic Surgeons National Congenital Heart Surgery Database Committee in association with Summit Medical Systems. Twenty-four centers joined the program at various dates of entry resulting in 18,894 enrolled patient records. This report compiled the relevant clinical features of 18 congenital heart categories over a 4-year period (1994–1997), which included 8,149 patient records. The data analyses are largely descriptive in character. Missing data points were described and not omitted in the analysis. Statistical analysis was not performed due to missing data points in some categories. Certain trends, however, could be identified and are discussed. The first Society of Thoracic Surgeons National Congenital Heart Surgery Database Report has succeeded in establishing a finite record that can be improved to establish universal national and international utility, risk stratification, and scholarly outcome analyses.

	Introduction

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
The first report of the Society of Thoracic Surgeons (STS) National Congenital Heart Surgery Database Committee in association with Summit Medical Systems was recently published [1]. Twenty-four centers (Appendix 1) joined the program at various dates of entry resulting in 18,894 enrolled patient records. The distribution of all primary diagnosis entries for the 18,894 enrolled patient records is listed in Appendix 2. This report compiled the relevant clinical features of 18 congenital heart categories over a 4-year period (1994–1997), which included 8,149 patient records. Emphasis was placed on the total number of patients for each category, the types of operations that were performed, the methods of myocardial protection, and five specific features of repair. Outcome data included operative death, complications, and length of stay (LOS). Outcome analyses were segregated for age or weight at operation where appropriate, which varied from diagnosis to diagnosis.

The data analyses were largely descriptive in character and identified missing data points, which were not entered by some participating centers. Rather than eliminate these patients from the analyses, the nonentries were listed as "missing data" to allow the reader to assess the relevance and importance of the data.

Similar to the STS adult cardiac surgery database, this report of the STS National Congenital Heart Surgery Database generated a massive amount of data. These data depicted many trends and were largely predictive of the established previous clinical reports from different centers. The analysis also demonstrated the strengths and weaknesses of a database, which, by necessity, limited the data input. On one hand, the four-page data form was readily available and concise; on the other hand, the information was limited and did not allow discriminating features that are necessary to establish risk stratification analysis. Unlike the STS adult cardiac surgery database, the Congenital Heart Surgery Database has numerous disease entities to analyze and by nature of the subspecialty, requires increased complexity in data analysis to produce meaningful risk stratification.

This inaugural effort of the STS National Congenital Heart Surgery Database Committee established the necessary foundation for future efforts to create an International Congenital Heart Surgery Database. The groundwork was created in areas of data organization for analysis. Important lessons learned from this initial database will shape the development of the next generation database, which will be computer based in cyberspace rather than limited by the size of four sheets of paper. This will allow for the collection of increased data input, the elimination or at least reduction of missing data points, and the generation of data specific enough to possess discriminating features necessary to establish risk stratification analysis. This next generation database will need to meet the dual goals of facilitating multiinstitutional data analysis, as well as providing data to support clinical programs, research, and teaching at individual institutions. The potential exists to create the first comprehensive international database for a medical subspecialty.

	Analysis of diagnoses

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
Atrial septal defect
Parameters for inclusion in the atrial septal defect (ASD) subgroup study population were repair of ASD as the primary procedure with a primary diagnostic field of a specific subtype of ASD. Records with other concomitant diagnoses were excluded from the study population except for those with associated left superior vena cava.

The distribution of isolated ASD by type in the STS study population reflects that described by other researchers [2]. More than 1,500 patients were reported; 90.7% secundum, 7.3% sinus venosus, 0.8% common atrium, and 1% residual type. Approximately 4 in 5 patients (81.8%; 1,132 of 1,384) with secundum ASD underwent operation with preoperative echocardiographic study only; 13% had preoperative echocardiographic and catheterization studies, and 1.7% had preoperative catheterization only; data were missing on 43 patients (3.1%). Similarly, approximately 3 of 4 patients (77%; 88 of 114) with sinus venosus ASD underwent operation with preoperative echocardiographic study only and 20% (23 of 114) had preoperative echocardiographic and catheterization studies; data were missing on 3 patients (2.6%). Although the number of patients with common atrium or residual ASD is small (28 patients), here too, 83% and 50% of the total common atrium and residual ASD patients, respectively, underwent operation with preoperative echocardiographic study only. This preference for echocardiography as the primary preoperative diagnostic tool was constant throughout the 4-year study period (1994–1997).

Median age at repair of all types of ASDs was 5.1 years; 25th and 75th percentiles were 3.0 and 11.5 years, respectively. Female-to-male ratio in the total ASD study population was 1.6:1. The ratio was 2:1 in the common atrium subgroup, 1.7:1 in both the secundum and residual subgroups, and 1.1:1 in the sinus venosus subgroup.

Cardiopulmonary bypass (CPB) was used for repair of ASD. As a measure of data validity, only 4.3% (66 of 1,526 patients) of ASD records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 68.4% (999 of 1,460 patients) used cross-clamping of the aorta, whereas 20.8% (303 of 1,460 patients) used induced fibrillation; 10.8% (158 of 1,460 patients) of the records were missing data on the specific technique used. Of the 999 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 54.2% (541 of 999 patients), crystalloid cardioplegia in 30.5% (305 of 999), and other cardioplegia in 5.7% (57 of 999 patients); data were missing in 9.6% (96 of 999) of the records. Closure of the ASD was by suture in 52.0% (794 of 1,526 patients) or by patch in 44.6% (681 of 1,526 patients); data fields were not completed in 51 (3.3%) records.

The complication incidence in isolated ASD closure is low. Of the 1,526 records in the study period, 1,425 (93.4%) reported no complications. There was only 1 operative death (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 0.07%. Of note, this death was in a patient with residual ASD diagnosis. Slightly < 1% of patients (15 of 1,526 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 10 patients (0.66%). Other major system complications were reported in a small number of patients: 2 major neurologic complications (0.13%), 9 pulmonary complications (0.59%), and 1 renal complication (0.07%). Eighty-three patients (5.44%) had complications other than those specifically mentioned.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data. From 1994 to 1997 the percentage of same-day operation patients increased from 54.3% in 1994 to 74.1% in 1995, 77.8% in 1996, and 79.9% in 1997. As a caveat, however, records with incomplete data were excluded; thus, 33% (507 of 1,526 patients) of the ASD study population records were not included in this analysis. The data available on preoperative LOS, however, support the trend toward same-day operation; here, missing dates account for a loss of only 99 of 1,526 records (6.5%).

Postoperative LOS data remained relatively constant over the study period. Mean postoperative LOS was 3.7 days and was 3.7, 3.4, 4.0, and 3.5 days in 1994 through 1997, respectively. Although the subgroup’s numbers are small, the LOS of patients with the residual ASD was 1 day longer on average.

Ventricular septal defect
Parameters for inclusion in the ventricular septal defect (VSD) subgroup study population were repair of VSD as the primary procedure with a primary diagnostic field of a specific subtype of VSD. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, patent ductus arteriosus, or left superior vena cava.

The distribution of isolated VSD by type in the STS study population reflects that described by other researchers [3]. In the 4-year study period, almost 1,200 patients were reported; 78.2% perimembranous, 8.8% conal, 5.6% muscular, 3.4% inlet (atrioventricular canal type), 3.1% multiple, and 0.9% residual type. More than one-half of the patients (52.2%; 477 of 914) with perimembranous VSD underwent operation with preoperative echocardiographic study only; 42.5% had preoperative echocardiographic and catheterization studies, and 1.7% had preoperative catheterization only; data were missing on 33 patients (3.6%). Although the number of patients reported was much smaller, it is of interest that patients with conal and inlet type VSDs were more likely to have only preoperative echocardiographic diagnostic studies, 62.1% (64 of 103 patients) and 57.5% (23 of 40 patients), respectively. For these two subgroups, 35% of the conal patients and 40% of the inlet patients had both preoperative echocardiographic and catheterization studies. One patient in the conal subgroup had a preoperative catheterization only; in the inlet subgroup, no patient had catheterization only. In the multiple, muscular, and residual subgroups, fewer patients underwent only preoperative echocardiographic study; 41.7% (15 of 36 patients) in the multiple subgroup, 35.4% (23 of 65) in the muscular subgroup, and 20% (2 of 10 patients) in the residual subgroup. Concomitantly, the percentage of patients in these groups who underwent both echocardiographic and catheterization studies preoperatively was higher; 52.8% (19 of 36 patients) in the multiple subgroup, 58.5% (38 of 65) in the muscular subgroup, and 80% (8 of 10 patients) in the residual subgroup. Of note, in 1994 43.1% (129 of 299 patients) had preoperative echocardiographic diagnostic study only. This increased to 53.1% (163 of 307) in 1995, 51.3% (160 of 312) in 1996, and finally, 60.8% (152 of 250 patients) in 1997. Concomitantly, preoperative echocardiographic and catheterization studies decreased from a high in 1994 of 49.5% (148 of 299 patients) to a low of 37.2% (93 of 250 patients) in 1997. The small number of patients over the study period who had preoperative catheterization study only makes trend analysis impossible.

Median age at repair of all types of VSDs was 0.69 years (8 months); 25th and 75th percentiles were 0.37 years (4 months) and 2.5 years, respectively. The female-to-male ratio in the total VSD study population and the perimembranous subgroup was 1:1. Men predominated in only one subgroup, the conal, where the male-to-female ratio was 1.5:1. Women predominated in the muscular, inlet, multiple, and residual subgroups, where the female-to-male ratio was 1.2:1, 1.9:1, 2:1, and 1.5:1, respectively.

Pulmonary artery band (PAB) placed at a previous operation was noted in 51 of the 1,168 records (4.4%). Within this group, 33.3% (12 of 36) were placed in patients with multiple VSDs, 24.6% (16 of 65) were placed in those with muscular VSDs, and 2.2% (20 of 914) were placed in patients with perimembranous defects. Only 1 patient (of 103; 1%) had a previous PAB placement in the conal VSD subgroup, and 2 patients (of 40; 5%) had band placement in the inlet VSD subgroup. The infrequent use of PAB is in keeping with other reports [4]. Cardiopulmonary bypass was used for repair of VSD. As a measure of data validity, only 8.9% (104 of 1,168) of VSD records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 99.1% (1,054 of 1,168) used cross-clamping of the aorta, whereas 0.2% (2 of 1,168) used induced fibrillation; 0.8% (8 of 1,168) of the records were missing data on the specific technique used.

Of the 1,064 patients with known CPB status, 142 (13.35%) had circulatory arrest. The percentage, in descending order, of circulatory arrest utilization per defect subgroup was 27.3% (6 of 27) in those patients with multiple VSDs, 20.8% (11 of 53) in those with muscular VSDs, 18.9% (7 of 37) in those with inlet type VSDs, 13.1% (111 of 844) in those with perimembranous defects, 11.1% (1 of 9) in those with residual defects, and 6.1% (6 of 99) in those with conal defects. Of note, the reported use of circulatory arrest declined over the 4-year study period, from a high of 53 patients (19.6%) reported in 1994, 41 (15.0%) in 1995, 34 (12.0%) in 1996, to a low of 14 patients (6.0%) in 1997. Whether this is a declining trend in use of circulatory arrest or reflects declining reporting on the part of users is subject to speculation. The percentage of use of circulatory arrest within two age groups (< 6 months of age/> 6 months of age) remains consistent over time, even with the decreasing number of patients reported, suggesting that this is a trend versus an artifact of missed reported cases. In 1994, 34 of 53 patients (64.1%) reported were < 6 months of age; in 1997, 9 of 14 patients reported (64.3%) were < 6 months of age; however, there was a drop in 1995 and 1996 to 58.5% and 55.9%, respectively, in patients reported < 6 months of age.

The subgroup of patients having circulatory arrest during VSD repair was further analyzed according to age at repair. Those < 6 months of age represented 60.6% (86 of 142) of all patients having circulatory arrest. The percentage for each subgroup, in decreasing order, within those patients < 6 months of age having circulatory arrest is, 85.7% (6 of 7 patients) with inlet-type VSDs, 83.3% (5 of 6) with conal VSD, 81.8% (9 of 11) with muscular VSD, 58.6% (65 of 111) with perimembranous VSD, and 16.7% (1 of 6 patients) with multiple VSDs. In the 56 patients > 6 months of age, 83.3% (5 of 6) had multiple VSDs, 41.4% (46 of 111) had perimembranous defects, and 18.2% (2 of 11) had muscular VSDs. There was 1 patient each in three of the subgroups, inlet, conal, and residual, representing 14.3%, 16.7%, and 100%, respectively, of the VSD population > 6 months of age having circulatory arrest.

Of the 1,054 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 62% (653 patients), crystalloid cardioplegia in 21.8% (230), and other cardioplegia in 7.3% (77 patients); data were missing in 8.9% (94) of the records. Closure of the VSD was by suture in 4.0% (47 of 1,168 patients) or by patch in 85.4% (998 of 1,168 patients); data fields were not completed in 123 records (10.5%). Patches were placed either transatrial (899 of 998, 90.1%), transpulmonary artery (55 of 998, 5.5%), transventricular (right or left) (42 of 998, 4.2%), or transaorta (2 of 998, 0.2%).

The complication incidence in isolated VSD closure is low. Of the 1,168 records in the study period, 1,001 (85.7%) reported no complications. There were 11 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 0.9%. Only two VSD subgroups had rates > 1%: multiple VSDs (3 of 36, 8.3%) and muscular VSDs (1 of 65, 1.5%). Slightly more than 3% (37 of 1,168) of patients underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 19 patients (1.6%). Other major system complications were reported in a small number of patients: 4 major neurologic complications (0.3%), 50 pulmonary complications (4.3%), and 3 renal complications (0.3%). In addition to or other than those major system or operative complications specifically mentioned, 122 patient records (10.4%) were marked as having "other" complications.

When the complication incidence is further analyzed by age, those < 6 months of age represent 36.3% of the population (424 of 1,168 patients). Of these, only 337 of the 1,001 (33.7%) reported no complications. Operative mortality was 1.65% in this subgroup (7 of 424) versus 0.5% (4 of 744) in those > 6 months of age. Rather than reporting absolute percentages within the < 6 months of age cohort, the proportion of complications in this age group relative to the entire VSD population reveals the impact that early age at repair has on incidence of complications. Within every category of complications analyzed (with the exception of renal complications), the < 6 months of age subgroup had incidences ranging from 50% to 70% of the complications noted for VSD repair for all ages. Seven of the total number of VSD operative deaths (of 11, 63.6%) were in this age group. Twenty of 37 patients (54%) were < 6 months of age and underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Ten of 19 patients (52.6%) were < 6 months of age and had complications attributable to infection. Other major system complications were reported in a small number of patients: 2 major neurologic complications (of 4 patients, 50%), 35 pulmonary complications (of 50, 70%) and 1 renal complication (of 3, 30%). In addition to or other than those major system or operative complications specifically mentioned, 62 patient records (of 122, 50%) were marked as having "other" complications.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data. From 1994 to 1997 the percentage of same-day operation patients increased from 43.5% in 1994 to 58.4% in 1995, 69.7% in 1996, and 69.5% in 1997. As a caveat, however, records with incomplete data were excluded; thus, 44.3% (518 of 1,168) of the VSD study population records were not included in this analysis. On the basis of 1,082 reports of same-day admission and operation, when further broken down into age subcategories, 158 of 384 patients were < 6 months of age (41.2%) and were reported to be admitted for same-day operation, versus 492 of 698 patients > 6 months of age (70.5%).

The data available on preoperative LOS, however, support the trend toward same-day operation; here, missing dates account for a loss of only 86 of 1,168 records (7.4%). When the data are further analyzed by age less or more than 6 months of age, for the 4-year study period, the mean preoperative LOS for patients < 6 months of age was 6.8 days versus 0.9 days for patients > 6 months of age. These younger patients are, by implication, admitted, presumably, for medical stabilization; their higher morbidity is reflected in their higher incidence of postoperative complications as discussed. It is beyond the scope of the database to comment on whether the higher incidence of complications may have had their root in preoperative morbid states or was truly secondary to the surgical repair.

Postoperative LOS data remained relatively constant over the study period. Mean postoperative LOS was 7.2 days and was 8.4, 6.6, 6.9, and 6.9 days in 1994 through 1997, respectively. Although some of the patient numbers of the subgroups are small, the mean postoperative LOS ranged from 5.2 days for the conal VSD subgroup (n = 98), to 9.6 days for the muscular VSD subgroup (n = 59). The mean postoperative LOS of inlet VSD patients was 6.7 days (n = 37), and was 7.2, 8.3, and 9.1 days for the perimembranous (n = 838), multiple (n = 30), and residual (n = 10) subgroups, respectively. The age of the patient at VSD repair had a profound impact on the postoperative LOS. For the 4-year study period, of 694 patients > 6 months of age, the mean postoperative LOS was 4.9 days; for those < 6 months of age, it was 11.4 days, more than double.

These data are largely confirmatory of recent publications on the subject [5, 6] and demonstrate the power of such a database when standards can assure purity of data input and retrieval.

Atrioventricular canal
Parameters for inclusion in the atrioventricular canal (AVC) subgroup study population were repair of AVC or ASD primum (with or without cleft mitral valve) as the primary procedure with a primary diagnostic field of a specific subtype of AVC or ASD primum. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, patent ductus arteriosus, or left superior vena cava.

In the 4-year study period, 590 patients were reported in the AVC defects group, 53.7% (317 of 590 patients) complete AVC, 36.6% (216 of 590) ASD primum, and 9.7% (57 of 590) intermediate AVC. One-half of the patients (50.5%, 160 of 317) with complete AVC underwent operation with preoperative echocardiographic study only; 43.8% (139 of 317) had preoperative echocardiographic and catheterization studies, and 1.3% (4 of 317) had preoperative catheterization only; data were missing on 14 patients (4.4%). Of the patients with ASD primum and intermediate AVC, 63.9% (138 of 216) and 54.4% (31 of 57), respectively, underwent operation with preoperative echocardiographic study only; 33.8% (73 of 216 patients) and 36.8% (21 of 57 patients), respectively, had preoperative echocardiographic and catheterization studies, and 2% (2 of 216 patients) and 3.5% (2 of 57 patients), respectively, had preoperative catheterization only; data were missing on 3 patients in each group (1.4% and 5.3%, respectively). Of note, the trend for preoperative echocardiographic study only in the ASD primum subgroup is strong, with the ratio of preoperative echocardiographic study only-to-echocardiogram + catheterization increasing from 1.1:1 in 1994, to 2.2:1 in 1995, to 1.9 in 1996, and finally, to 3:1 in 1997.

Median age at repair of all types of AVCs for the 4-year study period was 0.56 years (between 6 and 7 months); 25th and 75th percentiles were 0.37 years (4 months) and 1.5 years, respectively. Female-to-male ratio in the total AVC study population was 1.3:1; in the complete AVC subgroup, 1.1:1; in the ASD primum subgroup, 1.7:1; and in the intermediate AVC subgroup, 1:1.

Pulmonary artery band placed at a previous operation was noted in 10 of the 590 records (1.7%). Within this group, 7 were placed in patients with complete AVCs, and 3 were placed in patients with primum ASD. This is in keeping with the reports of other researchers, with primary repair the procedure of choice in the infant with AVC defect [7].

Cardiopulmonary bypass (CPB) was used for repair of AVC. As a measure of data validity, only 6.1% (36 of 590) of AVC records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 99.1% (549 of 554 patients) used cross-clamping of the aorta, whereas 0.2% (1 of 554) used induced fibrillation; 0.7% (4 of 554) of the records were missing data on the specific technique used.

Of the 549 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 69.2% (380 of 549 patients), crystalloid cardioplegia in 15.9% (87 of 549), and other cardioplegia in 8.2% (45 of 549); data were missing in 6.7% (37 of 549) of the records. Closure of complete AVC defect was by single-patch technique in 39.4% (125 of 317 patients) or with use of double patch in 56.2% (178 of 317 patients); data fields were not completed in 14 records (4.4%). Closure of intermediate AVC defect was by single-patch technique in 86% (49 of 57) or with use of double patch in 8.8% (5 of 57); data fields were not completed in 3 records (5.3%). Closure of ASD primum defect was by single-patch technique in 76% (164 of 216 patients); data fields were not completed in 52 records (24.7%). Double patch technique is inappropriate for ASD primum repair. We found it encouraging that double patch technique did not get recorded in the analysis.

Of the 590 AVC defect records in the study period, 459 (77.8%) reported no complications. There were 17 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 2.9% for the 4-year study period. Fourteen of the 17 deaths (82.4%) occurred in the complete AVC subgroup for a mortality rate of 4.4% (14 of 317 patients), 1 in the intermediate AVC subgroup for a mortality rate of 1.8% (of 57), and 2 in the ASD primum subgroup for a mortality rate of 0.9% (of 216).

Other specific complications and their incidence will be discussed in terms of each defect, not for the group as a whole. In addition, within each of the three subgroups (complete AVC, intermediate AVC, primum ASD), data were analyzed for the impact of patient age (< 6 months versus > 6 months) at the time of operation on outcome.

In the complete AVC subgroup, in patients < 6 months of age, no complications were recorded for 63.6% (133 of 209) of the patients; in those > 6 months of age, no complications were recorded for 81.5% (88 of 108). There were 11 operative deaths in the < 6 months of age complete AVC subgroup (11 of 209, 5.3%); in those > 6 months of age, 3 deaths occurred (3 of 108, 2.8%). Almost 7% (14 of 209, 6.7%) of patients < 6 months of age underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure; in those > 6 months, 8 of 108 (7.4%) underwent reoperation. Complications attributable to infection occurred in 9 patients < 6 months of age (9 of 209 patients, 4.3%); in those > 6 months of age, the incidence was 2.8% (3 of 108 patients). Other major system complications were also reported. There were two major neurologic complications in the < 6 months population segment (of 209, 1.0%), none in those > 6 months of age. Almost 40 patients in the < 6 months population segment had pulmonary complications (38 of 209, 18.2%); in those > 6 months of age, the incidence was 7.4% (8 of 108). Renal complications were noted in 3 of 209 (1.4%) patients in the < 6 months of age population segment, none in those > 6 months of age. In addition to or other than those major system or operative complications specifically mentioned, 57 patient records (of 209, 27.3%) were marked as having "other" complications in the < 6 months of age segment; only 13 of 108 (12%) records were so marked in those > 6 months of age.

In the intermediate AVC subgroup, in patients < 6 months of age, no complications were recorded for 75% (15 of 20) of the patients. In those > 6 months of age, no complications were recorded for 78.4% (29 of 37 patients); the only area where complications were noted in this age group was in "other" complications (discussed later). There was one operative death in the < 6 months of age intermediate AVC subgroup (of 20, 5%). Five percent (1 of 20) of patients < 6 months of age underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 2 patients < 6 months of age (of 20, 10%); there were no major neurologic complications, 2 of 20 (10%) had pulmonary complications, and 1 of 20 (5%) had renal complications. In addition to or other than those major system or operative complications specifically mentioned, 5 patient records (of 20, 25%) were marked as having "other" complications in the < 6 months of age segment; 8 of 37 (21.6%) records were so marked in those > 6 months of age.

In the primum ASD subgroup, in patients < 6 months of age, no complications were recorded for 70.4% (19 of 27) of the patients; in those > 6 months of age, no complications were recorded for 92.6% (175 of 189). There were two operative deaths in the < 6 months of age primum ASD subgroup (of 27, 7.4%); in those > 6 months of age, no deaths were reported. Eleven percent (3 of 27, 11.1%) of patients < 6 months of age underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure; in those > 6 months, 2 of 189 (1.1%) underwent reoperation. Complications attributable to infection occurred in 2 patients > 6 months of age (of 189, 1.1%). Other major system complications were also reported. There was one major neurologic complication in the < 6 months population segment (of 27, 3.7%), two in those > 6 months of age (of 189, 1.1%). Four patients in the < 6-month population segment had pulmonary complications (of 27, 14.8%); in those > 6 months of age, the incidence was 1.6% (3 of 189 patients). There were no renal complications noted in either age group. In addition to or other than those major system or operative complications specifically mentioned, 7 patient records (of 27, 25.9%) were marked as having "other" complications in the < 6 months of age segment; 11 of 189 (5.8%) records were so marked in those > 6 months of age.

In all three diagnostic subgroups, patients <6 months of age were at higher risk for mortality and morbidity.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data. From 1994 to 1997, 308 patient records indicated planned same-day surgical admission; the percentage of same-day operation patients increased from 40% in 1994 to 59.2% in 1995, 62.7% in 1996, and 64.2% in 1997 (percentages based on 545 records with admission and surgical date fields completed). Thus, 47.8% (282 of 590 patients) of the AVC defect study population records were not included in this analysis. When further broken down into age categories, 107 of 308 patients < 6 months of age (45.5%) were reported to be admitted for same-day operation versus 201 of 308 patients > 6 months of age (64.8%).

The data available on preoperative LOS, however, support the trend toward same-day operation; here, missing dates account for a loss of only 45 of 590 records (7.6%). The 50th percentile figures from 1994 to 1997 were 1 day, 1 day, 0 day, and 0 day, respectively, for the < 6-month age group; for the > 6-month age group, they were 1 day, 0 day, 0 day, 0 day. When the data are further analyzed by age less or more than 6 months of age, for the 4-year study period, the mean preoperative LOS for patients < 6 months of age was 5.0 days, versus 1.6 days for patients > 6 months of age. These younger patients are, by implication, admitted, presumably, for medical stabilization; their higher morbidity is reflected in their higher incidence of postoperative complications as discussed. It is beyond the scope of the database to comment on whether the higher incidence of complications may have had their root in preoperative morbid states or were truly secondary to the surgical repair. When the data are analyzed by defect subgroup, the mean preoperative LOS over the 4-year study period for the complete AVC subgroup for whom admission dates were available was 4.3 days, for the intermediate AVC subgroup mean preoperative LOS was 1.3 days, and for the ASD primum subgroup it was 1.7 days.

Postoperative LOS data remained relatively constant over the study period. Mean postoperative LOS for the AVC defect population was 9.7 days and was 11.4, 8.4, 9.8, and 9.6 days in 1994 through 1997, respectively. The mean postoperative LOS ranged from 6 days for the ASD primum subgroup (n = 198), to 7.9 days for the intermediate AVC subgroup (n = 55), to 12.7 days for the complete AVC subgroup (n = 278). The age of the patient at AVC repair had a profound impact on the postoperative LOS. For the 4-year study period, of 309 AVC defect population patients > 6 months of age, the mean postoperative LOS was 6.6 days; for those < 6 months of age, it was 14 days, more than double. Within the defect subgroups, those < 6 months of age with ASD primum had a mean postoperative LOS of 11 days; those > 6 months of age, 5.3 days. Those < 6 months of age with intermediate AVC had a mean postoperative LOS of 11.8 days; those > 6 months of age, 5.8 days. Those < 6 months of age with complete AVC had a mean postoperative LOS of 14.6 days; those > 6 months of age, 9.3 days.

These data are reflective of some of the more favorable reports in the literature concerning outcome after repair of AVC [7, 8]. The less or more than 6-month of age data are interesting and should help to establish future risk stratification criteria and cost projections.

Truncus arteriosus
Inclusion criteria were primary diagnosis of truncus arteriosus (truncus), with indicated subtype. Allowable concomitant diagnoses included ASD, patent ductus arteriosus (PDA), significant truncal insufficiency, and left superior vena cava.

The distribution of the Van Praagh and Van Praagh [9] class subtypes was similar to recent reports [10, 11]. Of the 193 reported cases (1994–1997), 63.7% (123 of 193 patients) were class A1, 23.8% (46 of 193) were class A2, 6.8% (13 of 193) were class A3, and 5.7% (11 of 193 patients) were class A4. Significant truncal insufficiency according to class was present in 14 patients (11.4%), 2 (4.3%), 3 (23.1%), and 1 patient (9.1%), respectively. The principal methods of preoperative diagnosis underwent a change over the study period indicating a greater reliance on echocardiography. The percentages of patients who had echocardiography, without cardiac catheterization over the study period were 29.5% in 1994, 32.7% in 1995, 47.5% in 1996, and 52.6% in 1997. Conversely, the percentage of patients undergoing catheterization and echocardiography decreases over the study period, 56.8%, 53.9%, 47.5%, and 44.7%, respectively. The number of patients who had only a catheterization study was small (2.1%); the number of missing entries for method of diagnosis was 13 (6.7%).

Median age at repair for all patients was 1 month; 25th and 75th percentiles were 2 weeks and 5.5 months, respectively. The female-to-male ratio for all patient types was 1.2:1.

Cardiopulmonary bypass was used for repair of truncus. As a measure of data validity, 10.9% (21 of 193) of truncus records had missing CPB data (time and CPB run). Further analysis of those records with CPB data showed that 90.1% (155 of 172) used aortic cross-clamping and 1.2% (2 of 172) used induced fibrillation; 8.7% (15 of 172) of the records were missing data on the specific technique used. Of the 155 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 63.2% (98 of 155 patients), crystalloid cardioplegia was used in 17.4% (27 of 155), other types of cardioplegia were used in 13.6% (21 of 155), and missing entries comprised 5.8% (9 of 155 patients). The data involving the features and repair were disappointing. The standard repair for truncus involves a valved right ventricular to pulmonary artery conduit. Only 52.9% (102 of 193 patients) had recorded conduit data. Interestingly, 100% of the 102 patient records that included conduit data subgrouped the conduits into homograft types (98 of 102) or bioprosthetic (non-human) types (4 of 102).

The complication rate for truncus repair is relatively high. Of the 193 records during the study period, 121 (62.7%) reported no complications. There were 22 operative deaths (11.4%), 4 of whom had severe truncal insufficiency (death within hospitalization or within 30 days of surgical procedure). More than 16% (31 of 193 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications due to infection occurred in 6 patients (3.1%), to pulmonary problems in 29 (15.0%), and to other conditions not otherwise specified in 51 patient records (26.4%). Operative mortality for truncus was further analyzed by age at operation, which included: < 1 month of age, 1–3 months of age, and > 3 months of age. The percentage mortality was 14.6% (14 of 96 patients), 17.1% (7 of 41), and 1.8% (1 of 56 patients), respectively. It would seem from these data that patients do better when the operation is performed after 3 months of age. These data, however, do not reflect the inherent mortality during the waiting period before the operation is performed. As expected, the mortality for type A1 truncus (8.1%) is about 50% less than the mortality for types A2, A3, and A4, which were 15.2%, 23.1%, and 18.2%, respectively. These expected findings reflect the greater hemodynamic problems associated with the most complex pathoanatomy.

Postoperative LOS data reflected the complexity and severity of the disease. Mean postoperative LOS for all truncus patients for the 4-year study period was 17.6 days (25th and 75th percentiles were 7 and 20 days, respectively), and was 13.6, 17.0, 18.0, and 22.3 days in 1994 through 1997, respectively. When separated into subgroups by age, the average LOS for < 1 month of age, 1–3 months of age, and > 3 months of age were 20.9, 20.9, and 10.8 days, respectively. These trends reflect recent clinical reviews [10, 11].

Aortopulmonary window
Inclusion criteria for aortopulmonary window (AP window) were primary diagnosis of AP window (all types). Allowable concomitant diagnoses were ASD, VSD, PDA, and left superior vena cava. The number of enrolled patient records over the 4-year study period (1994–1997) was small (14 patients) and reflected on the rarity of the disease [12]. Diagnosis was by echocardiography only in 71.4% (10 of 14), cardiac catheterization only in 0%, and by both cardiac catheterization and echocardiography in 7.1% (1 of 14 patients). There were 3 (21.4%) missing data points. Female-to-male ratio was 1:1. Features of repair reflected mainstream techniques. The use of CPB (time of CPB run) was recorded in 12 of 14 patients (85.7%) and therefore, either missing or not used in 2 patients (14.3%). Missing data also confounded the use of cross-clamp in 9 of 12 patients (75.0%) and cardioplegia (blood cardioplegia in 6 and crystalloid cardioplegia in 2 patients), which was recorded in 8 of 9 patients (88.9%) whose records indicated that aortic cross-clamping was used.

Operative mortality (death within hospitalization or within 30 days of surgical procedure) over the 4-year study period (1994–1997) was zero (0%), whereas 71.4% (10 of 14) of patient records recorded no complications. Postoperative LOS was recorded in 13 of 14 patients and was a mean of 9.9 days, 25th and 75th percentiles 5 and 7 days, respectively.

The small number of recorded cases and paucity of relevant data acquisition made this analysis difficult and almost irrelevant. For instance, because of the small numbers, we did not segregate those patients who had associated VSD. We could not examine the clinical outcomes in relation to the different types of AP window and to the various coronary artery anomalies because there were no designated categories on the database. Interestingly, we decided against too much detail in low volume diagnoses such as these when the database was being established. In light of these results with AP window, one could argue that more data rather than less data, especially in the low volume diagnoses, are indicated to more effectively analyze these relatively rare cases.

Anomalous origin of coronary artery from the pulmonary artery
Parameters for inclusion in the anomalous origin of coronary artery from the pulmonary artery (ACAPA) study group population were primary diagnosis of ACAPA. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, PDA, mitral regurgitation, or left superior vena cava. The number of patients with ACAPA was surprisingly low based on the number of participating centers and enrolled patients. Forty-six patient records were identified over the 4-year study period; 44 had anomalous left main coronary artery arising from the pulmonary artery (ALCAPA) and 2 had anomalous right coronary artery arising from the pulmonary artery (ARCAPA). No trends were noted over time in the method of diagnosis. The principal method of diagnosis was by echocardiography alone in 36% (16 of 46 patients), by cardiac catheterization alone in 2.2% (1 of 46), and by echocardiography and cardiac catheterization in 58.7% (27 of 46 patients); missing data for diagnosis occurred in 2 patients (4.4%). Recent publications [13, 14] have extolled the benefits and feasibility of echocardiography-only diagnosis, which was not noted in this study. The average age at operation for ALCAPA during the 4-year period was 4.5 months (25th and 75th percentiles were 2.5 and 9.5 months, respectively). The female-to-male ratio for ACAPA was 3.2:1. The number of patients having same-day operation over the study period was small (8 of 46 patients) (20.5%).

Cardiopulmonary bypass was used for repair of ACAPA. As a measure of data validity, 6.5% (3 of 46 ACAPA records) had missing CPB data (time of CPB run). Further analysis of these records with CPB data revealed 100% (43 of 43) used cross-clamping of the aorta for the repair. No patient record showed the use of induced fibrillation and there were no missing methods. Of the 43 patient records with cross-clamp times specified for the 4-year period, blood cardioplegia was used in 69.8% (30 of 43 patients), crystalloid cardioplegia in 20.9% (9 of 43), and other cardioplegia in 7.0% (3 of 43 patients). There was only one missing method. Because the operative repair for ACAPA involves procedures on the coronary arteries, an analysis on cardioplegia delivery (antegrade versus retrograde) was made. Antegrade, retrograde, and both antegrade as well as retrograde cardioplegia was used in 60.5% (26 of 43), 4.7% (2 of 43), and 27.9% (12 of 43 patient records), respectively. Cardioplegia delivery method was missing in 6.9% (3 of 43).

This database did not provide a comprehensive list of treatment options for ACAPA such as ligation, Takuchi operation, reimplantation, internal thoracic artery coronary bypass, or subclavian to coronary anastomosis. The choices for the features of ACAPA repair in this database were ligation, reimplantation/coronary artery bypass grafting, and concomitant extracorporeal membrane oxygenation. Reimplantation/coronary artery bypass grafting, which includes all methods of coronary bypass and reimplantation, was used in the greater majority of patients (89.1%, 41 of 46 patients), with simple ligation used in 4.4% (2 of 46). Missing methods occurred in 6.5% (3 of 46 patients). Concomitant extracorporeal membrane oxygenation was used in 6.5% (3 of 46 patient records) as a method of perioperative left ventricular support.

The complication incidence in ACAPA reflected the problems associated with left ventricular dysfunction, which is so prevalent in this patient population. Of the 46 patient records in the study period, 28 (60.9%) reported no complications. There were five operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality of 10.9% for the 4-year study period. About 13% (6 of 46 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Associated major postoperative complications occurred in a relatively large percentage of patients over the study period: infection 10.9% (5 of 46 patients), pulmonary 23.9% (11 of 46), renal 2.2% (1 of 46), and "other" 30.4% (14 of 46 patients).

When the complication incidence is further analyzed by age, those < 1 year of age (37 of 46 patients) had all the mortality (5 deaths, 13.5%), whereas those > 1 year of age (9 of 46) had no mortality. These data are expected and reflect the difference in the collateral coronary circulation that inexplicably develops in some patients providing adequate myocardial perfusion and fails to develop in others resulting in poor myocardial perfusion, ventricular dysfunction, and mitral regurgitation. Some investigators [15] have conveniently labeled the adequate collateral group as the "adult type" and the poor collateral group as the "infantile type," which of course explains these data. In any case, this type of nomenclature becomes moot in light of newer and more effective diagnostic studies that identify most patients in infancy, thereby making the diagnostic dilemma irrelevant as early operation is recommended in virtually all patients.

Postoperative LOS data remained relatively long and constant over the study period. Mean postoperative LOS was 19.3 days (25th and 75th percentiles, 5 and 19 days, respectively) and was 18.3, 20.5, 11.8 and 22.0, respectively over the 4-year study period (1994–1997). When LOS data were segregated for age the mean LOS for patients < 1 year of age was 23.2 days; the mean LOS for patients > 1 year of age was 5.6 days.

The number of patients in this group was small. In addition, the limited scope of the input data sheet did not allow therapeutic segregation into the various types of revascularization strategies. Those debating future database schemes will have to consider these issues.

Coarctation of the aorta
Parameters for inclusion in the coarctation of the aorta (CoA) subgroup study population were repair of CoA as the primary procedure with a primary diagnosis of CoA (all types). Records with other concomitant diagnoses were excluded from the study population except for those with associated PDA or left superior vena cava.

The distribution of isolated CoA over the study period showed a decline in numbers reported from a high of 168 patients in 1994 to 132 patients in 1997. This most likely reflects the increase over this same time period in the use of catheter dilation of CoA. Of the 591 patients reported, 28.4% (168 patients) were in 1994, 28.3% (167) in 1995, 21% (124) in 1996, and 22.3% (132 patients) in 1997. The CoA population was split into three age groups for some analyses: those patients < 1 month of age, those 1 month to 1 year of age, and those > 1 year of age at the time of surgical repair. There were 32.7% (193 patients) < 1 month, 24.2% (143) 1 month to 1 year, and 43.1% (255 patients) > 1 year of age at operation.

Approximately 71% of CoA patients (421 of 591, 71.2%) underwent operation with preoperative echocardiographic study only; 19.6% (116 of 591) had preoperative echocardiographic and catheterization studies, and 1.4% had preoperative catheterization only. Data were missing on 46 patients (7.8%). This preference for echocardiography as the primary preoperative diagnostic tool was constant (range was 73.2% in 1994 to 69.7% in 1997) throughout the 4-year study period (1994–1997).

Median age at repair of CoA was 0.39 years (between 4 and 5 months of age); 25th and 75th percentiles were 0.04 (2 weeks) and 5.4 years, respectively. Female-to-male ratio in the total CoA study population was 1:1.8.

Cardiopulmonary bypass was used in repair of coarctation in 6.3% (37 of 591 records), and not used in 93.7% (554 of 591). Of the 37 patients with known CPB use, the aorta was cross-clamped in 33 (89.2%); data were missing in the remaining 4 patients (10.8%).

In the population of patients with known cross-clamp time (33), the records are marked by incomplete field entries. Eighteen of the 33 (54.5%) records had missing field entries. Blood cardioplegia was used in 9 patients (27.3%); crystalloid cardioplegia in 6 (18.2%). It is not clear from this data whether there were patients with unrecorded associated VSD. This would explain the use of CPB and cardioplegia use. The difficulty, however, is that this explanation does not segregate those patients who had partial CPB to repair isolated coarctation.

Incidence of type of CoA repair (subclavian flap, synthetic patch, resection with end-to-end anastomosis, or interposition graft) was analyzed with regard to patient age. In patients < 1 month of age, resection with end-to-end anastomosis was used in 70% of the records (135 of 193 patients), subclavian flap repair in 14% (27), and synthetic patch repair in 0.5% (1 patient); data were missing in 30 records (15.5%). In patients 1 month to 1 year of age, resection with end-to-end anastomosis was used in 73.4% of the records (105 of 143 patients), subclavian flap repair in 10.5% (15 of 143), synthetic patch repair in 4.2% (6 of 143), and interposition graft in 0.7% (1 of 143 patients); data were missing in 16 records (11.2%). In patients > 1 year of age, resection with end-to-end anastomosis was used in 56.5% of the records (144 of 255), synthetic patch repair in 19.2% (49 of 255), interposition graft in 7% (18 of 255), and subclavian flap repair in 5.5% (14 of 255); data were missing in 30 records (11.8%).

The complication incidence in isolated CoA repair is low. Of the 591 records in the study period, 549 (92.9%) reported no complications. There were four operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 0.7%. Of note, three of the four deaths (75%) occurred in patients < 1 month of age and the other occurred in a patient 1 month to 1 year of age. Operative complications were present in 9 patients (of 591, 1.5%) who then underwent reoperation for related problems. Complications attributable to infection occurred in 7 patients (1.2%). Other major system complications were reported in a small number of patients: 3 major neurologic complications (0.5%), 13 pulmonary complications (2.2%), and 2 renal complications (0.3%). Twenty-nine patients (4.9%) had complications other than those specifically mentioned. Other than major neurologic complications, one incident was reported in each age subgroup; the majority of complications occurred in the < 1 month of age subgroup: 7 of 9 operative complications (77.8%), 4 of 7 infections (57.1%), 11 of 13 pulmonary complications (84.6%), 2 of 2 renal complications (100%), and 20 of 29 other complications (69%). No complications were noted in 85.5% (165 of 193 patients) of the < 1 month of age subgroup, versus that noted in the 1 month to 1 year and > 1 year of age subgroups, 93% (133 of 143) and 98.4% (251 of 255), respectively.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data, with 47.7% (256 of 537 patients) admitted for same-day operation over the 4-year study period. Records with incomplete data were excluded; thus, 9.1% (54 of 591) of the CoA study population records were not included in this analysis. From 1994 to 1997 the percentage of same-day operation patients increased from 35.9% (52 of 145) in 1994 to 53.5% (85 of 159) in 1995, 52.2% (59 of 113) in 1996, and 50% (60 of 120) in 1997. The data available on preoperative LOS support the trend toward same-day operation; 50th percentile preoperative LOS in 1994 was 1 day, in 1995 and 1996, 50th percentile preoperative LOS was 0 days, and in 1997 50% percentile preoperative LOS was 0.5 days.

Postoperative LOS data remained relatively constant over the study period. Mean postoperative LOS for all CoA patients for the 4-year study period was 6.2 days, and was 6.4, 6.8, 5.4, and 5.7 days in 1994 through 1997, respectively. When separated into subgroups by age, the postoperative LOS of < 1 month of age patients for the 4-year period was 10.1 days; for the 1-month to 1-year age group, 5.4 days, and for the > 1-year age group, 3.8 days. The individual year postoperative LOS did not differ in great degree from the 4-year study period figure for all years except for the LOS for the < 1-month age subgroup. Here, the mean stay decreased from 10.2 days in 1994, increased to 12.4 days in 1995, and decreased again to 7.9 days in 1996 and 8.3 days in 1997.

These trends reflect recent reports that emphasize extended end-to-end anastomosis and favorable survival statistics [16, 17].

Patent ductus arteriosus
Parameters for inclusion in the patent ductus arteriosus (PDA) subgroup study population were division or ligation of PDA as the primary procedure with a primary diagnosis of PDA. Records with other concomitant diagnoses were excluded from the study population except for those with associated left superior vena cava.

The distribution of isolated PDA over the study period showed a decline in numbers reported from a high of 350 patients in 1994 to 230 patients in 1997. This most likely reflects the increase over this same time period in the use of catheter occlusion of PDA [18]. The PDA population was split into two groups for some analyses: those patients weighing < 2 kg at operation and those weighing > 2 kg at operation. There were 53.9% (632 of 1,172 patients) weighing < 2 kg and 46.1% (540 of 1,172 patients) weighing > 2 kg at operation. More than 1,100 patients were reported, with 29.9% (350 of 1,172) in 1994, 27.4% (321 of 1,172) in 1995, 23.1% (271 of 1,172) in 1996, and 19.6% (230 of 1,172 patients) in 1997. Approximately 86% of PDA patients (1,007 of 1,172, 85.9%) underwent operation with preoperative echocardiographic study only; 3.9% (46 of 1,172) had preoperative echocardiographic and catheterization studies, and 0.1% had preoperative catheterization only. Of note, in patients weighing > 2 kg, diagnosis was by echocardiography only in 89.3% of patients (482 of 540), and by echocardiography and catheterization study in 7.2% of patients (39 of 540). Data were missing on 118 patients (10.1%); of these, 100 were patients who weighed < 2 kg at operation. This preference for echocardiography as the primary preoperative diagnostic tool was constant throughout the 4-year study period (1994–1997).

Median age at repair of PDA was 0.27 years (3 months of age); 25th and 75th percentiles were 0.03 years (11 days) and 2.8 years, respectively. Female-to-male ratio in the total PDA study population was 1.4:1.

The complication incidence in isolated PDA closure is low. Of the 1,172 records in the study period, 1,114 (95%) reported no complications. There were 18 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 1.5%. Of note, 15 of the 18 deaths (83.3%) occurred in patients < 2 kg in weight at the time of operation. Slightly < 1% (11 of 1,172 patients) underwent reoperation for related complications. Complications attributable to infection occurred in 8 patients (0.7%). Other major system complications were reported in a small number of patients: 1 major neurologic complication (0.1%), and 33 pulmonary complications (2.8%); 29 of the 33 incidents were noted in patients weighing < 2 kg at operation (87.9%). There were no reported renal complications. Twenty-four patients (2.1%) had complications other than those specifically mentioned.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data [19]. From 1994 to 1997 the percentage of same-day operation patients increased from 46.4% (135 of 291 patients) in 1994 to 55.9% (156 of 279) in 1995, 52.4% (111 of 212) in 1996, and 54.4% (87 of 160 patients) in 1997. As a caveat, however, records with incomplete data were excluded; thus, 19.6% (230 of 1,172 patients) of the PDA study population records were not included in this analysis; most of the missing data were from the < 2-kg patient population (205 of 230 patients). The data available on preoperative LOS, however, support the trend toward same-day operation; 50th percentile preoperative LOS in 1994 was 1 day, in 1995 through 1997, 50th percentile preoperative LOS was 0 days. When patient weight is a factor, mean preoperative LOS differs in the two groups. In the > 2-kg group, mean preoperative LOS for the 4-year study period is 1.4 days; in the < 2-kg group, 4.4 days, reflecting the multiple medical problems faced by these low-weight infants.

Postoperative LOS data remained relatively constant over the study period. Mean postoperative LOS for all PDA patients for the 4-year study period was 14.9 days, and was 15.6, 12.4, 13.6, and 19.6 days in 1994 through 1997, respectively. When separated into groups by weight, the < 2-kg patients postoperative LOS for the 4-year period was 25.9 days; for the > 2-kg group, 5.9 days. Of import, the < 2-kg group had 218 missing dates so that only 414 of 632 eligible patients’ LOS were analyzed. The > 2-kg group had only 27 records (of 540) with missing dates. The individual year postoperative LOS did not differ in great degree from the 4-year study period figure for all years except 1997. In 1997, mean postoperative LOS for patients > 2 kg was 7.8 days; for patients < 2 kg, 40.2 days. Again, missing data for 65 patients in the < 2-kg group makes difficult any reliable comparison between the two groups. However, the generally higher postoperative LOS seen in 1997 may reflect the selection out of the patient pool of the straightforward, uncomplicated PDA patient to undergo catheter occlusion rather than surgical division or ligation.

Pulmonary atresia and stenosis with intact ventricular septum
Parameters for inclusion in the pulmonary atresia and stenosis with intact ventricular septum (PA/PS-IVS) study population were all types of primary diagnosis of PA/PS-IVS. Allowable concomitant diagnoses included ASD, PDA, right ventricular to pulmonary outflow tract obstruction, coronary fistula-congenital, right heart lesions (tricuspid stenosis and Ebstein’s anomaly), and left superior vena cava.

The database allowed for this category to be further segregated into patients with PA-IVS and PS-IVS. During the 4-year study period (1994–1997) 284 patient records were entered; 165 patients (58.1%) had PA-IVS and 119 (41.9%) had PS-IVS. The number of patient record entries per year remained relatively constant over the study period (1994–1997), 78, 76, 82, and 48, respectively. Almost two-thirds of the patients (65.5%, 186 of 284 patients) with PA/PS-IVS underwent operation with preoperative echocardiography and cardiac catheterization; 26.4% (75 of 284 patients) had preoperative echocardiographic studies only and 3.5% (10 of 284 patients) had preoperative catheterization only; data were missing on 10 patients (4.6%). It is more than passing interest to note that many of these reported patients were prospectively enrolled in the Congenital Heart Surgeons Society Study on PA/PS-IVS, which required, among other things, preoperative diagnostic assessment by cardiac catheterization and echocardiography [20]. This accounts for the relatively high incidence of preoperative cardiac catheterization. After this study and largely because of it, better echocardiographic assessment was established leading to the ongoing trend toward echocardiography and away from cardiac catheterization for diagnosis in the neonate and the infant.

The way that this database is constructed, we know that CPB was used when the time of CPB was recorded (CPB run). If the CPB run was omitted, other indirect means to determine CPB use are present such as the use of fibrillation/cross-clamp cardioplegia, and obvious open heart procedures. Despite these clues "missing data" were assumed in up to 10% of the other analyses. The difficulty with PA/PS-IVS is that there are bona fide techniques for treatment using CPB and not using CPB. In this database we know by the "time and CPB run" that CPB was used. If this patient is not entered, then we must assume that either CPB was used and not entered making this a "missing method" or that CPB was intentionally not used. We cannot tell the difference between "CPB not used" or "missing data." With this caveat in mind, CPB was used in 71.8% (204 of 284 patients) and CPB not used or missing data was present in 28.2% (80 of 284 patients). In those patients who had CPB recorded, aortic cross-clamping was used in 60.8% (124 of 204 patients), involved fibrillation in 6.4% (13 of 204), and "missing method" in 32.8% (67 of 204 patients). In those patients who had aortic cross-clamping, blood cardioplegia was used in 76 of 124 patients (61.3%), crystalloid cardioplegia in 20 of 124 (16.1%), "other" cardioplegia in 13 of 124 (10.5%), and missing method in 15 of 124 patients (12.1%). Any attempt to segregate the operative procedures for PA/PS-IVS must take into consideration that there are numerous (more than 10) operative strategies to treat this disease. We chose to center on five treatment strategies that we believed would capture the greater majority of patients. Of the 284 patients in this group, 66 (23.2%) had pulmonary valvotomy with or without chamber or vessel patch augmentation with CPB (VO); 23 of these had PA-IVS and 43 had PS-IVS. It makes sense that most patients in this group had PS-IVS, as these are the types of patients who have the greatest chance of success without a systemic to pulmonary artery (S-P) shunt. A minimum of patients (6 of 284, 2.1%) had pulmonary valvotomy with or without chamber or vessel patch augmentation without CPB. We anticipated that more patients would have pulmonary valvotomy with or without chamber or vessel patch augmentation and an S-P shunt using CPB (VO + S) or not using CPB (VC + S). This turned out not to be the case as the analysis discovered only 7 of 284 patients (2.5%) that had VO + S and only 9 of 284 patients (3.2%) that had VC + S. The shunt only group (S) was predictive of other reports [20] indicating that 67 of 284 patients (23.6%) had S and most of these patients (57) predictably had PA-IVS, whereas only 10 patients had PS-IVS. The most disconcerting results in this analysis were the large number of "other" procedures, 129 of 284 (45.4%). To have such a large number of patients fall into categories other than the five that were included is not consistent with previous reports [20]. One has to conclude that the large number of therapeutic choices may have significantly confounded the data. Future database schemes have to account for this obvious problem.

The complication incidence for all the procedures was relatively low for this group of patients. Of the 284 records in this study, 226 (79.6%) reported no complications. There were 11 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 3.9%. The mortality rate for PA-IVS was slightly higher than the PS-IVS group, 8 of 284 (4.9%) and 3 of 284 (2.5%), respectively. Slightly more than 6% (18 of 284 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 4 patients (1.4%). Other major system complications were reported in a small number of patients: 1 neurologic complication (.4%), 21 pulmonary complications (7.4%), and 2 renal complications (0.7%). In addition to or other than those major system or operative complications specifically mentioned, 38 patient records (13.4%) were recorded as having "other" complications. The postoperative LOS data revealed a mean stay of 9.6 days (25th and 75th percentiles, 4 and 10 days, respectively).

Future database schemes will have to consider even more therapeutic strategies that are emerging from transcatheter techniques. A comprehensive database will be able to accurately segregate the various therapeutic procedures that should help clinicians treat this very difficult group of patients.

Pulmonary atresia with VSD
Parameters for inclusion in the Pulmonary Atresia with VSD (PA-VSD) group included primary diagnosis of PA-VSD or major AP collateral arteries (MAPCAs) with patch closure, or both entities. Records with other concomitant diagnoses were excluded from the study population except for those with associated VSD and left superior vena cava. This analysis by no means assesses the complexity of some of these patients who may undergo multiple S-P shunts or unifocalization procedures before coming to biventricular repair. Therefore, we chose to offer a "snap shot" of those patients who were deemed reparable regardless of the number of prerepair procedures. We also chose to segregate those patients by age: those who had biventricular repair < 6 months of age (8 of 55 patients), those who had biventricular repair at > 6 months of age but < 2 years of age (24 of 55 patients), and those who had biventricular repair at > 2 years of age (23 of 55 patients). We acknowledge that this could be a biased comparison because the mortality associated with the multiple unifocalization/shunt procedures leading up to biventricular repair in the older age groups does not enter into the calculation. Consequently, one would expect a higher mortality in the < 6 months of age group in which patients infrequently undergo shunting procedures.

Patients were segregated into two groups; 25 of 55 patients (45.5%) had biventricular repair for PA-VSD and 30 of 55 patients (54.5%) had biventricular repair for PA-VSD-MAPCAs over the 4-year study period (1994–1997). One would expect a preponderant reliance on cardiac catheterization in this study group and was confirmed by the data. Most patients 45 of 55 (81.8%) underwent diagnostic cardiac catheterization and echocardiography, whereas only 2 of 55 patients (3.7%) and 7 of 55 patients (12.7%) underwent only cardiac catheterization or echocardiography, respectively; there was a missing entry in 1 of 55 patients (1.8%).

Median age at repair for both groups was 1.7 years, (25th and 75th percentiles 0.7 and 3.3 years, respectively). Of some interest is that the PA-VSD group underwent biventricular repair earlier than the PA-VSD-MAPCAs group, 1.2 years (25th and 75th percentiles 0.8 and 1.9 years, respectively) and 2.6 years (25th and 75th percentiles 0.7 and 4.2 years, respectively). The female-to-male ratio was 1.1:1.

Cardiopulmonary bypass data (time of CPB run) was missing in only 2 of 55 (3.6%) of patient records; 53 of 55 (96.4%) had recorded CPB entries. Of these 53 patient records, all (100%) had cross-clamp entries, whereas 49 had cardioplegia entries; 41 of 53 patients (77.4%) used blood cardioplegia, 3 of 53 (5.7%) used crystalloid cardioplegia, 5 of 53 (9.4%) used "other" forms of cardioplegia, and 4 of 53 patients (7.5%) were the missing entries.

We tried to capture the features of the biventricular repair by noting which patients had previous S-P shunt takedown, the use of an extracardiac conduit (which we assumed was necessary in almost all patients), and the type of conduit used (homograft versus other types of bioprostheses). Although there were some "missing entries," the data seemed to confirm literature reports [21, 22]. S-P shunt takedown at the time of biventricular repair was recorded in 15 of 55 patients (27.3%); one might expect a higher percentage of S-P shunt takedown in light of the median age at repair (1.7 years). Valved conduits, 40 of 55 patients (77.7%), were preferred; no patient record (0%) indicated that a nonvalved conduit was used (nonvalved conduit was a choice in the database), and 15 of 55 patients (27.3%) had missing conduit data. Of those 40 patient records indicating conduit data, 34 of 40 (85.0%) had homografts and 6 of 40 (15.0%) had other bioprosthetic valved conduits.

Operative mortality (death within hospitalization or within 30 days of surgical procedure) was not insignificant, 6 of 55 (10.9%), with only 35 of 55 (63.6%) of patient records reporting no complications. Slightly more than 18% (10 of 55) of patients underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 1 patient (1.8%), to neurologic events in 1 patient (1.8%), to pulmonary complications in 11 patients (20.0%), to renal complications in 2 patients (3.6%), and due to "other" complications in 14 of 55 patients (25.5%). In spite of the aforementioned disclaimers, we noted mortality rates of 37.5% (3 of 8 patients), 4.2% (1 of 24), and 8.7% (2 of 23 patients) in the < 6-month-old group, the > 6-month-old but < 2-year-old group, and the > 2-year-old group, respectively. Whether the mortality in the < 6-month-old group could improve if the biventricular repair were to be delayed is left to speculation and further analysis of more detailed data. The postoperative LOS in survivors was similar in both groups with a total mean LOS time of 12.6 days, 25th and 75th percentiles 5 and 14 days, respectively. If stratified by age the < 6-month-old group median LOS was 25.4 days. The median LOS for the older age groups were 12.2 and 10.1 days, respectively.

Aside from the obvious gross measurements that were made in this group over the 4-year study period (1994–1997), one is struck by the need for prerepair surgical as well as medical therapy to properly evaluate this very complex group of patients. Future data gathering schemes will require more information to discriminate issues relating to age at operation, prerepair therapies that recruit pulmonary arteries (including transcatheter techniques), and prerepair mortality rates.

Tetralogy of Fallot
Parameters for inclusion in the tetralogy of Fallot (TOF) subgroup study population were primary diagnosis of TOF or TOF with associated absent pulmonary valve, anomalous coronary artery, or AVC and repair of TOF as the primary current procedure with VSD closure. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, PDA, left superior vena cava, or right ventricle-to-pulmonary outflow tract obstruction.

The distribution of TOF by type in the STS study population reflects that described by other researchers [23]. In the 4-year study period, 773 patients were reported, 88.7% with no associated lesion, 6.2% with absent pulmonary valve, 2.2% with anomalous coronary artery, and 2.9% with AVC. Children with TOF commonly undergo preoperative echocardiographic and cardiac catheterization studies, ranging from a high of 76.5% (13 of 17 patients) in the TOF with anomalous coronary artery subgroup, 63.6% (14 of 22) in the TOF with AVC subgroup, 61.4% (421 of 686) in the straightforward TOF group, to a low of 58.3% (28 of 48 patients) in the subgroup with absent pulmonary valve; data were missing on 28 patients (3.6%). The subgroup with anomalous coronary artery was least likely to have preoperative echocardiographic study only (11.8%, 2 of 17 patients), whereas those with absent pulmonary valve (13 of 49 patients, 27.1%), AVC (7 of 22, 31.8%), or straightforward TOF (211 of 686 patients, 30.8%) were more likely to have preoperative echocardiographic study only. There appears to be a trend toward increasing use of preoperative echocardiographic study only. Over the 4-year study period, the number of preoperative "echo only" cases remained relatively constant, ranging from 53 in 1994, 59 in both 1995 and 1996, and 62 in 1997; respective percentages, however, increased over this period from 23.9% to 29.4%, 32.1%, and 37.4%. The number of patients reported having both echocardiographic and catheterization studies preoperatively, however, decreased substantially over the same period from a high of 142 patients in 1994 to 125 patients in 1995, 112 in 1996, and 97 patients in 1997, representing 64%, 62.2%, 60.9%, and 58.4% of the patients, respectively.

Median age at repair of all types of TOF was 0.73 years (9 months); 25th and 75th percentiles were 0.31 years (4 months) and 1.4 years, respectively. The anomalous coronary artery subgroup was operated on at an older age; median age at repair was 2.9 years with 25th and 75th percentiles 1.7 and 4.7 years, respectively. The AVC subgroup patients were also operated on a later age; median age at repair was 1.1 years with 25th and 75th percentiles 0.7 years (8 months) and 1.8 years, respectively.

Female-to-male ratio in the total TOF study population was 0.7:1. Males predominated in all groups, with the male-to-female ratio 1.3:1 in the straightforward TOF subgroup, 1.2:1 in the absent pulmonary valve and AVC subgroups, and 1.1:1 in the anomalous coronary artery subgroup (the subgroup with the smallest number of patients, 17).

Cardiopulmonary bypass was used for repair of TOF. As a measure of data validity, 16.8% (130 of 773) of TOF records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 94.4% (607 of 643 patients) used cross-clamping of the aorta, whereas 0.2% (1 of 643 patients) used induced fibrillation; 5.4% (35 of 643) of the records were missing data on the specific technique used.

Of the 608 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 65.1% (395 of 608 patients), crystalloid cardioplegia in 21.1% (128 of 608), and other cardioplegia in 6.4% (39 of 608 patients); data were missing in 7.4% (45 of 608) of the records.

Features of repair that were analyzed included incidence of takedown of previously placed shunts, placement of a transannular patch, and insertion of bioprosthetic pulmonary valve. Of the total TOF population over the 4-year study period, 69 of 773 (8.9%) had takedown of a previously placed shunt (58 in the straightforward TOF subgroup, 5 in the absent pulmonary valve subgroup, and 3 each in the anomalous coronary artery and AVC subgroups); 235 of 773 patients (30.4%) had transannular patch placement (220 in the straightforward subgroup, 8 in the absent pulmonary valve subgroup, 2 in the anomalous coronary artery subgroup, and 5 in the AVC subgroup); and 8 of 773 patients (1.0%) had a bioprosthetic pulmonary valve placed (5 in the straightforward TOF subgroup and 3 in the absent pulmonary valve subgroup).

Conduit placement as a feature of TOF repair was also analyzed. Over the 4-year study period, 38 valved and 0 nonvalved conduits were reported placed during the course of TOF repair (38 of 773 patients, 4.9%). Of these 38 conduits, 35 were homograft (92.1%) and 3 bioprosthetic (7.9%). Conduits were reported placed in only 14 patients in the straightforward TOF subgroup (14 of 686, 2.0%), 20 of 48 (41.7%) in the absent pulmonary valve subgroup, and 4 of 22 patients (18.2%) in the AVC subgroup; no conduits were reported placed in the anomalous coronary artery subgroup.

The complication incidence in TOF repair is mixed. Of the 773 records in the study period, 588 (76.1%) reported no complications. This ranged from a high of 78.6% (539 of 686 patients) in the straightforward TOF subgroup, to 64.7% (11 of 17) in the anomalous coronary artery subgroup, 63.6% (14 of 22) in the AVC subgroup, to a low of 50% (24 of 48 patients) in the absent pulmonary valve subgroup. There were 22 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 2.9% for the 4-year study period. The highest mortality rates were in the absent pulmonary valve (5 of 48 patients, 10.4%) and AVC (2 of 22 patients, 9.1%) subgroups. The straightforward TOF subgroup had 15 deaths (15 of 686 patients, 2.2%). Fifty-two of 773 patients (6.7%) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Pulmonary complications were reported in 53 patients (6.9%). Not unexpectedly, the absent pulmonary valve subgroup reported 14 of 53 patients or 26.4%. Thirty-two patients (4.1%) had complications attributable to infection. Other major system complications were reported in a small number of patients: 3 major neurologic complications (0.4%), and 8 renal complications (1.0%). In addition to or other than those major system or operative complications specifically mentioned, 127 patient records (16.4%) were marked as having "other" complications.

The complication incidence in the TOF population was further segmented and analyzed by age, < 1 month (75 of 773 patients, 9.7%), 1–6 months (193 of 773, 25%), and > 6 months of age (505 of 773 patients, 65.4%). Rather than reporting absolute percentages within the < 1-month of age cohort, the proportion of complications in this age group relative to the entire TOF population reveals the impact that early age at repair has on incidence of complications. Within most of the categories analyzed (with the exception of neurologic and "other") the < 1-month cohort, representing < 10% of the TOF population, had incidences ranging from 21% to 26% of the complications noted for TOF repair for all ages. Almost 23% of the mortalities reported (5 of 22, 22.7%) occurred in the < 1-month segment, 31.8% (7 of 22) in the 1–6-month segment, and 46.9% (10 of 22) in the > 6 months of age segment. In the < 1-month of age segment, 14 of 53 patients (26.4%) had pulmonary complications noted, 8 of 32 patients (25%) had complications attributable to infection, 2 of 8 patients (25%) had renal complications, and 11 of 52 patients (21.2%) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. However, in addition to or other than those major system or operative complications specifically mentioned, only 16 patient records (16 of 127, 12.6%) were marked as having "other" complications.

The trend toward increasing use of same-day operation in the pediatric population was reflected by the STS data. From 1994 to 1997 the percentage of same-day operation patients increased from 30.3% in 1994 to 44.9% in 1995, 52.6% in 1996, and 58.4% in 1997. As a caveat, however, records with incomplete admission/discharge data were excluded (59 of 773 patients, 7.6%). On the basis of 329 reports of same-day admission and operation (of 721, 45.6%), when further broken down into population age segments, 7 (2.1%) were < 1 month of age, 57 of 329 (17.3%) were between 1 and 6 months of age, and 265 of 329 (80.5%) were > 6 months of age. Analysis by diagnosis reveals that patients with straightforward TOF and those with associated AVC were admitted for same-day operation for approximately half the admissions (304 of 639 or 47.6% and 11 of 21 or 53.4%, respectively). Of note, patients with TOF and associated absent pulmonary valve had a mean preoperative LOS of 8.69 days, not unexpected in these children [24].

The data available on preoperative LOS show that age at admission impacts the preoperative LOS. For the 4-year study period, the mean preoperative LOS for patients < 1 month of age was 4.2 days, versus 3.0 days for patients 1–6 months of age and 1.1 days for patients > 6 months of age. These younger patients are, by implication, admitted, presumably, for medical stabilization; their higher morbidity is reflected in their higher incidence of postoperative complications as discussed. Again, it is beyond the scope of the database to comment on whether the higher incidence of complications may have had their root in preoperative morbid states or were truly secondary to the surgical repair.

Postoperative LOS data were analyzed for the 4-year study period with respect to patient age at operation. Patients < 1 month of age (60 of 700 with reported discharge dates) had mean postoperative stays of 15.5 days (25th percentile 5 days, 50th percentile 9.5 days, 75th percentile 21.5 days). Those patients between 1–6 months of age (169 of 700) had mean postoperative stays of 9.8 days (25th percentile 5 days, 50th percentile 7 days, 75th percentile 10 days). Patients > 6 months of age (471 of 700) had mean postoperative stays of 7.8 days (25th percentile 4 days, 50th percentile 6 days, 75th percentile 9 days). Obviously, age at operation has an impact on postoperative LOS.

Length of stay data were also analyzed by type of TOF. Mean postoperative LOS for straightforward TOF patients (625 of 700 reported patients with complete admission data fields) and those with associated anomalous coronary artery (16 of 700) were 8.3 days (25th percentile 4 days, 50th percentile 6 days, 75th percentile 9 days) and 7.6 days (25th percentile 4 days, 50th percentile 5 days, 75th percentile 7.5 days). For those patients with associated AVC (19 of 700), mean postoperative LOS was 10.5 days (25th percentile 5 days, 50th percentile 8 days, 75th percentile 15 days). Last, those patients with associated absent pulmonary valve (40 of 700) had mean postoperative LOS of 18.8 days (25th percentile 8.5 days, 50th percentile 14 days, 75th percentile 25.5 days). These patients’ compromised respiratory status impacts their postoperative course.

Total anomalous pulmonary venous connection
Inclusion criteria for total anomalous pulmonary venous connection (TAPVC) study population were primary diagnosis of TAPVC, with the indicated subtype. Allowable concomitant diagnoses were ASD and PDA.

The distribution of TAPVC by type in this study population reflects that described by other researchers [25, 26]. In the 4-year study period (1994–1997) 168 patients were reported, 58.9% (99) had supracardiac (type I) lesions, 11.9% (20) had intracardiac (type II) lesions, 19.7% (33) had infracardiac (type III) lesions, and 9.5% (16) had mixed (type IV) lesions.

The method of diagnosis reflects the growing reliance on echocardiography alone in these often very sick patients. Echocardiography, as the only means of diagnosis, was used in 65.5% (110 of 168) of patients; cardiac catheterization only was used in 0.6% (1 of 168 patients) and cardiac catheterization as well as echocardiography in 29.7% (50 of 168 patients). Missing methods were noted in 4.2% (7 of 168 patients). These trends were constant over the entire study period.

Median age at repair for all TAPVC subtypes was 0.04 years (about 2 weeks of age), 25th and 75th percentiles 0.008 years (3 days) and 0.2 years (2.5 months), respectively. The median age at repair for types I, II, III, and IV were 0.05 years (3 weeks), 0.14 years (7.5 weeks), 0.02 years (7 days), and 0.03 years (11 days), respectively. As might be expected, the type III TAPVC patients (infracardiac), which have the highest incidence of pulmonary venous obstruction, were operated on at the earliest age; median age was 0.019 years (7 days), 25th and 75th percentiles 0.003 years (1 day) and 0.06 years (3 weeks), respectively. The male-to-female ratio was 2.2:1. Males predominated in all groups.

Cardiopulmonary bypass was used for repair of TAPVC. As a measure of data validity, 8.9% (15 of 168 patients) of TAPVC records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 97.3% (149 of 168 patients) used cross-clamping of the aorta, whereas 0.7% (1 of 168 patients) used induced fibrillation; 2.0% (3 of 168 patients) of the records were missing data on the specific technique used.

Of the 149 patient records with cross-clamp times specified for the 4-year period, blood cardioplegia was used in 66.4% (99), crystalloid cardioplegia in 16.1% (24), other types of cardioplegia in 7.4% (11), and missing entries comprised 10.1% (15) of patient records.

We assumed that complete repair of TAPVC is by direct anastomosis of the confluence of pulmonary veins to the left atrium. We did not sample for such variable as: intentional patency of the vertical vein, intentional patency of the patent foramen ovale, use of deep hypothermia and circulatory arrest, or whether the chamber entry was through the left or right atrium. Also, we did not sample for preoperative extracorporeal membrane oxygenation, which is occasionally inadvertently used in the event of late or mistaken diagnosis. All of these parameters are discoverable within this database and can be analyzed by proper inquiry programs, assuming that the input mechanisms were accurate and complete. We simply chose to limit the scope of our initial inquiry.

The complication incidence in TAPVC repair reflects recent surgical series [25, 26]. Of the 168 records in the study period, 125 (74.4%) reported no complications; this trend was constant over the 4-year study period (1994–1997). There were 14 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 8.3%. The highest mortality rates were in the type III (infracardiac) and type IV (mixed) subgroups, which were 12.1% (4 of 33 patients) and 18.8% (3 of 16 patients), respectively. Types I (supracardiac) and II (intracardiac) had lower mortality rates of 6.1% (6 of 99 patients) and 5.0% (1 of 20 patients), respectively. Sixteen of 168 patients (9.5%) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Pulmonary complications were reported in 21 of 168 patients (12.5%); complications relating to infection, neurologic events, and renal problems were present in 2.4% (4 patients), 0.6% (1), and 2.4% (4 patients), respectively. Other complications not mentioned comprised 18.5% (31 of 168 patients) of the study population.

Postoperative mean LOS for the entire cohort of surviving patients was 11.2 days, 25th and 75th percentiles 6 and 13 days, respectively. When analyzed for specific subtype the postoperative mean LOS for type I TAPVC was 12.4 days, 25th and 75th percentiles 6 and 14 days, respectively. The postoperative mean LOS for type II TAPVC was 8.6 days, 25th and 75th percentiles 4 and 10 days, respectively. The postoperative mean LOS for type III TAPVC was 10.5 days, 25th and 75th percentiles 8 and 13 days, respectively. The postoperative mean LOS for type IV TAPVC was 8.2 days, 25th and 75th percentiles 6 and 13 days, respectively.

This analysis of TAPVC shows the favorable aspects of this database. The results are similar to reported surgical procedures and the number of "missing entries" was relatively small. This may be related to a number of factors. The diagnostic choices within the database were clear and included synonyms for each subtype. The operation for each of the subtypes is the same with only small variations between them, making the analyses easier to perform. The age at presentation for operation is largely within the first year of life and therefore, represents a homogenous age group for analysis. Future databases may require more detailed entries and long-term outcome data to answer some of the more complex questions concerning this disease.

Transposition of the great arteries
Parameters for inclusion in the transposition of the great arteries (TGA) subgroup study population were repair of TGA as the primary procedure with a primary diagnostic field of a specific subtype of TGA. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, VSD, PDA, PS-IVS, right ventricle-to-pulmonary outflow tract obstruction (chamber or vessel), pulmonary insufficiency, or left superior vena cava.

The distribution of TGA by type in the STS study population reflects that described by other researchers [27, 28]. In the 4-year study period, 486 patients were reported, 56.2% with IVS, 27% with VSD, and 16.9% with VSD and PS. Over the 4-year study period, an increasing percentage of patients underwent operation with preoperative echocardiographic study only. This was noted especially in the TGA groups with IVS and VSD without associated PS. In 1994, 23.1% (18 of 78 patients) and 22.2% (6 of 27 patients) had preoperative echocardiographic study only in the IVS and VSD without associated PS groups, respectively. By 1997, this had increased to 41.9% (18 of 43 patients) and 46.2% (12 of 26 patients), respectively. Over the 4-year study period, 62.3% (170 of 273 patients) of the IVS subgroup underwent preoperative echocardiographic and catheterization study, as did 58.0% (76 of 131 patients) of the VSD subgroup, and 70.7% (58 of 82 patients) of the VSD with associated PS subgroup. During this same period, 32.6% (89 of 273 patients) of the IVS subgroup underwent preoperative echocardiographic study only, as did 33.6% (44 of 131 patients) of the VSD subgroup, and 24.4% (20 of 82 patients) of the VSD with associated PS subgroup. The number of patients with preoperative catheterization study only was very small. Data were missing on 19 patients (3.9%).

Median age at repair of TGA for all patients reported over the study period was 0.022 years (8 days); 25th and 75th percentiles were 0.014 years (5 days) and 0.12 years (6 weeks), respectively. Male-to-female ratio in the total TGA study population was 2.3:1.

Over the 4-year study period, arterial switch procedure was performed in 355 of 486 patients (73.1%), Rastelli procedure in 34 (7.0%), and Mustard/Senning procedure in 11 patients (2.3%); other procedure or data were missing in 87 patients (17.9%). As would be anticipated, in the TGA/VSD/PS subgroup, Rastelli procedure was performed in 39.0% (32 of 82 patients), whereas in the TGA/IVS and TGA/VSD subgroups, arterial switch was performed in 87.2% (238 of 273 patients) and 84.7% (111 of 132 patients), respectively.

Takedown of a previously placed PAB at the time of TGA repair was noted in 11 of the 486 records (2.3%). Six patients had PAB takedown before repair in the TGA/IVS subgroup; 4 patients before arterial switch procedure and 2 before a missing/other procedure. Four patients had PAB takedown before repair in the TGA/VSD subgroup; 2 before arterial switch and 2 before a missing/other procedure. One patient had PAB takedown before Rastelli in the TGA/VSD/PS subgroup.

Takedown of a previously placed shunt at the time of TGA repair was noted in 22 of the 486 records (4.4%). Six patients had shunt takedown before repair in the TGA/IVS subgroup; 3 patients before arterial switch procedure, 1 patient before Mustard/Senning, and 2 before a missing/other procedure. Two patients had shunt takedown before arterial switch in the TGA/VSD subgroup. Fourteen patients had shunt takedown before TGA repair in the TGA/VSD/PS subgroup; 1 patient before arterial switch, 12 patients before Rastelli procedure, and 1 before a missing/other procedure.

Takedown of both PAB and shunt at the time of TGA repair was noted in 6 of the 486 records (1.2%). Four patients had shunt and PAB takedown before repair in the TGA/IVS subgroup; 3 patients before arterial switch procedure and 1 before a missing/other procedure. One patient had shunt and PAB takedown before arterial switch in the TGA/VSD subgroup. One patient had shunt and PAB takedown before Rastelli in the TGA/VSD/PS subgroup.

Two patients had takedown of a Mustard/Senning procedure at the time of TGA repair in the TGA/VSD subgroup, 1 before arterial switch and 1 before a missing/other procedure.

Cardiopulmonary bypass was used for repair of TGA. As a measure of data validity, only 11.1% (54 of 486 patients) of TGA records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 97.9% (423 of 486 patients) used cross-clamping of the aorta, whereas 0.5% (2 of 486 patients) used induced fibrillation; 1.6% (7 of 486 patients) of the records were missing data on the specific technique used.

Of the 423 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 60.3% (255 patients), crystalloid cardioplegia in 22.7% (96), and other cardioplegia in 9.5% (40 patients); data were missing in 7.6% (32 of 423 patients) of the records.

The complication incidence for TGA repair was analyzed for the total population reported, as well as by type of repair. Of the 486 records in the study period, 333 (68.5%) reported no complications. The TGA/VSD/PS group had the lowest incidence of complications with 84.2% (69 of 82 patients) reporting no complications, followed by the TGA/IVS group (68.1% or 186 of 273 patients) and the TGA/VSD group (59.5% or 78 of 131 patients). There were 30 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 6.2%. This rate was constant over each subgroup: 6.2% (17 of 273 patients) in TGA/IVS, 6.1% (8 of 131) in TGA/VSD, and 6.1% (5 of 82 patients) in TGA/VSD/PS. Paralleling this finding, the TGA/VSD/PS subgroup reported only 7.3% (6 of 82 patients) operative complications with regard to reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure; those in the TGA/IVS group reported 15.4% (42 of 273 patients) and in the TGA/VSD group, 20.6% (27 of 131 patients). Complications attributable to infection occurred in 19 patients (3.9%), 9 reported patients in the TGA/VSD subgroup and 10 in the TGA/IVS subgroup. Other major system complications were reported in a small number of patients: 7 major neurologic complications (1.4%; 4 in the TGA/VSD subgroup and 3 in the TGA/IVS subgroup), 49 pulmonary complications (10.1%; 34 in the TGA/IVS subgroup and 15 in the TGA/VSD subgroup), and 9 renal complications (1.9%; 4 each in the TGA/IVS and TGA/VSD subgroups and 1 in the TGA/VSD/PS subgroup). In addition to or other than those major system or operative complications specifically mentioned, 98 patient records (20.2%) were marked as having "other" complications: 51 of 273 patients (18.7%) in the TGA/IVS subgroup, 35 of 131 (26.7%) in the TGA/VSD subgroup, and 12 of 82 patients (14.6%) in the TGA/VSD/PS subgroup.

When the complication incidence is further analyzed by type of operation, those patients undergoing arterial switch for repair of TGA had a mortality rate of 6.2% (22 deaths in 355 patients, 14 in the TGA/IVS subgroup, 7 in the TGA/VSD subgroup, and 1 in the TGA/VSD/PS subgroup). No complications were reported in 228 of 355 patients (64.2%) undergoing arterial switch. Operative complications with reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure were reported in 65 of 355 patients (18.3%); there were 38 incidents in the TGA/IVS subgroup, 25 in the TGA/VSD subgroup, and 2 in the TGA/VSD/PS subgroup. Complications attributable to infection occurred in 17 of 355 patients (4.8%), 8 reported in the TGA/IVS subgroup and 9 in the TGA/VSD subgroup. Other major system complications were reported in a small number of patients: 6 major neurologic complications (1.7%; 2 in the TGA/IVS subgroup and 4 in the TGA/VSD subgroup), 41 pulmonary complications (11.6%; 26 in the TGA/IVS subgroup and 15 in the TGA/VSD subgroup), and 7 renal complications (2.0%; 3 in the TGA/IVS and 4 in the TGA/VSD subgroups). In addition to or other than those major system or operative complications specifically mentioned, 79 patient records (22.3%) were marked as having "other" complications (46 in the TGA/IVS subgroup, 30 in the TGA/VSD subgroup, and 3 in the TGA/VSD/PS subgroup).

Data for the Mustard/Senning and Rastelli procedure subgroups are available; however, the small number of patients reported in each group makes interpretation difficult. Eleven patients underwent Mustard/Senning procedure; however, complete data were available on only 10 patients for analysis of complications and mortality. There were 3 operative deaths, for a mortality rate of 30%; 2 of the deaths occurred in the TGA/IVS subgroup and 1 in the TGA/VSD subgroup. Six of 10 patients (60%) reported no complications. There were 3 pulmonary, 1 infection, 1 major neurologic, 1 renal, and 3 other complications reported in this subgroup. Thirty-four patients underwent Rastelli procedure; however, complete data were available for only 33 patients. There were 2 operative deaths (in the TGA/VSD/PS subgroup), for a mortality rate of 6.1%. No complications were reported in 28 of 33 patients (84.9%). There were 3 operative, 1 renal, and 4 other complications reported in the Rastelli subgroup.

The trend toward increasing use of same-day operation in the pediatric population was not evident in the data for this diagnostic subgroup, most likely because of the early age at operation, as reported earlier. These patients are admitted shortly after birth and undergo relevant diagnostic procedures before operation, which is generally done at the same admission. The mean preoperative LOS for all TGA subgroups over the 4-year study period was 4.4 days; 50th percentile was 3 days, 25th and 75th percentiles were 1 and 6 days, respectively. From 1994 to 1997 the percentage of same-day operation patients increased from 12.1% in 1994, decreased to 10.5% in 1995, increased to 16.2% in 1996, and to 22.7% in 1997. The TGA/VSD/PS subgroup had the greatest number of same-day operation admissions, 25 (of a possible population of 82); however, 14 patients had missing relevant admission, surgical, or discharge dates and were excluded. Of the total TGA study population records 17.1% (83 of 486 patients) were not included in this analysis. In the TGA/IVS subgroup, those patients most likely to be operated early in the neonatal period, had the lowest number of same-day operation admissions reported, 16 of a possible 273 patients (48 patients were excluded because of missing dates). The TGA/VSD subgroup had 17 same-day operation admissions reported out of a possible 131 patients (21 patients were excluded).

Mean postoperative LOS for the 4-year study period was 13.9 days, and was 14.9, 12.5, 15.9, and 12.30 days in 1994 through 1997, respectively. The 50th percentile for postoperative LOS for the 4-year study period was 9 days, and was 10, 9, 9.5, and 9.5 days for 1994 through 1997, respectively. The 25th percentile for postoperative LOS for the 4-year study period was 7 days, and was 7, 7, 6, and 6 days for 1994 through 1997, respectively. The 75th percentile for postoperative LOS for the 4-year study period was 15 days, and was 15, 14, 16.5, and 14 days for 1994 through 1997, respectively. The TGA/VSD/PS subgroup had the shortest postoperative LOS. The mean postoperative LOS for this subgroup for the 4-year study period was 9.4 days, with a 5-, 7-, and 10-day LOS for the 25th, 50th, and 75th percentiles, respectively. The TGA/IVS subgroup’s mean postoperative LOS for the 4-year study period was 14.7 days, with a 7-, 10-, and 16-day LOS for the 25th, 50th, and 75th percentiles, respectively. The TGA/VSD subgroup’s mean postoperative LOS for the 4-year study period was 15.0 days, with a 7-, 10-, and 14-day LOS for the 25th, 50th, and 75th percentiles, respectively.

These data are largely consistent with literature reports [27, 28] and show a clear trend away from the atrial baffle operations and preference for the arterial switch operation. Length of stay is still quite long for all these patients and may reflect the noncardiac problems of neonatal life. In general this database will allow for a comprehensive risk stratification analysis assuming that missing data points are minimized. Notable omissions are information regarding abnormal coronary artery anatomy and newer operations for TGA-VSD-PS.

Aortic stenosis and left ventricular outflow tract obstruction
Parameters for inclusion in the aortic stenosis and left ventricular outflow tract obstruction (AS-LVOTO) subgroup study population were repair of AS-LVOTO as the primary procedure with a primary diagnostic field of a specific subtype of AS-LVOTO. Records with other concomitant diagnoses were excluded from the study population except for those with associated ASD, PDA, or left superior vena cava.

More than 400 patients were reported, 38.8% (160 of 412 patients) valvular, 9.2% (38 of 412) supravalvular, 48.6% (200 of 412) subvalvular, and 3.4% (14 of 412 patients) LV outflow obstruction type. Approximately 30% (49 of 160 patients) with valvular AS underwent operation with preoperative echocardiography study only; 58.7% (94 of 160 patients) had preoperative echocardiographic and catheterization studies, and 8.1% (13 of 160 patients) had preoperative catheterization only; data were missing on 4 patients (2.5%). Approximately 21% (8 of 38 patients) with supravalvular AS type underwent operation with preoperative echocardiographic study only and 71% (27 of 38 patients) had preoperative echocardiographic and catheterization studies, 5.2% (2 of 38) had preoperative catheterization only; data were missing on 1 patient (2.6%). For those patients with subvalvular AS type, 44.5% (89 of 200 patients) underwent echocardiographic study only, 51.5% (103 of 200 patients) underwent echocardiographic and catheterization studies, 1.5% (3 of 200 patients) underwent catheterization only; data were missing on 5 patients (2.5%). In the last group, patients with LV outflow obstruction, 14.3% (2 of 14) had echocardiographic studies only, 78.6% (11 of 14) had echocardiographic and catheterization studies, 0% had catheterization studies only, and data were missing on 1 patient (7.1%). There were no large trends away from or toward any diagnostic methods during the study period, 1994–1997.

Median age at repair of all types of AS was 7.6 years; 25th and 75th percentiles were 3.6 and 14.3 years, respectively. Male-to-female ratio was 1.7:1.

Cardiopulmonary bypass was used for repair of AS. As a measure of data validity, only 3.2% (13 of 412 patients) of AS had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 99.5% (397 of 399 patients) used cross-clamping, 0.50% (2 of 399 patients) of the records were missing data on the specific technique used. Of the 397 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 74.1% (294 of 397 patients) and crystalloid cardioplegia in 13.9% (55 of 397 patients); and other cardioplegia in 5.0% (20 of 397 patients); data were missing in 7.0% (28 of 397 patients) of the records. No major trends were noted in the types of cardioplegia that were used during the study period.

The STS Summit database allowed for a large number of procedure choices, which reflected the numerous surgical preferences that were used in this study population. Of the 160 patients with valvular stenosis, 4 (2.5%) had intraoperative dilatation, 47 (29.4%) had aortic valvuloplasty, 83 (51.2%) had aortic valve replacement, 14 (8.8%) had an annular enlarging procedure with aortic valve replacement, and 25 patients (15.6%) were marked as having "other" procedures. The sum of the percentages is > 100% owing to the possibility of more than one procedure for each patient. Of the 83 patients who had aortic valve replacement, 6 of 83 (7.2%) had a mechanical valve, 11 of 83 (13.2%) had a bioprosthetic valve, 56 of 83 (67.5%) had pulmonary autograft replacement (Ross procedure), and 10 of 83 patients (12.1%) were recorded as having "other or unknown" valve replacements. These trends are consistent with the extant literature [29], which documents the increasing application of the Ross procedure in these patients. Of the 14 patients who had annular enlarging operations in conjunction with aortic valve replacement, 4 patients (28.6%) had mechanical valves, 1 (7.1%) had a bioprosthetic valve, 6 (42.9%) had a pulmonary autograft, and 3 patients (21.4%) had "other or unknown" valves.

Patients with supravalvular aortic stenosis had either patch augmentation (26 of 38, 68.4%) or other procedures (12 of 38, 31.6%), which could include some of the newer aortoplasty procedures that do not use patch material.

Of the 200 patients who underwent procedures for subvalvular aortic stenosis, 169 (84.5%) had some form of muscle resection, whereas 31 (15.5%) had some other type of procedure. No patient had mitral valve replacement for this disease process during the study period.

Of the 412 records in the study period, 338 (82.0%) reported no complications. There were 6 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 1.5%. Approximately 4.1% (17 of 412 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications attributable to infection occurred in 9 patients (2.2%). Other system complications were reported as follows: 5 neurologic complications (1.2%), 5 pulmonary complications (1.2%), and 1 renal complication (0.2%). Fifty-seven patients (13.8%) had complications other than those specifically mentioned.

The mean postoperative LOS for the entire cohort of patients was 5.8 days, 25th and 75th percentiles were 3 and 6 days, respectively. The mean postoperative LOS was greater for the valvular group, 8.0 days, compared to the means of the supravalvular, subvalvular, and LV outflow obstruction groups, which were 4.8, 4.6, and 5.3 days, respectively.

These data do not reflect the trend away from aortic valvuloplasty and toward percutaneous aortic valve balloon dilatation, which has been increasingly reported [30]. Future comprehensive database schemes will require inclusion of invasive catheter techniques and longitudinal follow-up data if intelligent decisions are to be made based on the extant, therapeutic modalities.

Hypoplastic left heart syndrome
Inclusion criteria in the hypoplastic left heart syndrome (HLHS) study population were primary diagnosis of HLHS and cardiac transplantation, Norwood procedure, bidirectional Glenn shunt/Hemi-Fontan, or RA/caval to PA connection (Fontan/Kreutzer). Allowable concomitant diagnoses included ASD, PDA, coronary fistula-congenital, mitral stenosis, CoA, pulmonary venous stenosis, and left superior vena cava. Patients were stratified by the type of operation that was performed during the 1994–1997 study period. Of the 348 patients in this study, 16.9% (59 of 348) had cardiac transplant, 53.7% (187 of 348) had the Norwood procedure, 20.9% (73 of 348) had bidirectional Glenn shunt/Hemi-Fontan, and 8.3% (29 of 348) had RA or caval to PA connection (Fontan/Kreutzer). Approximately 68% (40 of 59 patients) with cardiac transplant underwent operation with preoperative echocardiographic study only; 22% (13 of 59 patients) had preoperative echocardiographic and catheterization studies, and 0% had preoperative catheterization only; data were missing on 6 patients (10.1%). Similar preoperative diagnostic evaluation was noted for patients undergoing the Norwood procedure; 64.7% (121 of 187 patients) had preoperative echocardiographic study only, 28.8% (54 of 187 patients) had preoperative echocardiographic and catheterization studies, 1% (2 patients) had preoperative catheterization only; data were missing on 10 patients (5.3%). In the bidirectional Glenn shunt/Hemi-Fontan patients, 19.1% (14 of 73 patients) underwent preoperative echocardiographic study only, 75.3% (55 of 73) underwent preoperative echocardiographic and catheterization studies, 1.3% (1 of 73 patients) had catheterization only; data were missing on 3 patients (10.3%). For the last group of patients, RA or caval to PA (Fontan/Kreutzer), 44.8% (13 of 29 patients) underwent preoperative echocardiographic study; 44.8% (13 of 29 patients) had preoperative echocardiographic and catheterization, 0% had catheterization only; data were missing in 3 patients (10.3%).

Median age at repair for the Norwood operation was 0.016 years (6 days), 25th and 75th percentiles were 4 and 10 days, respectively. As a comparison, the median age at cardiac transplantation was 0.120 years (about 6.5 weeks), 25th and 75th percentiles were 3 weeks and 3.5 months, respectively. Median age for bidirectional Glenn or Hemi-Fontan was 0.548 years (about 6.5 months), 25th and 75th percentiles, 0.482 years (about 5.5 months) and 0.720 years (about 8.5 months), respectively. The median age for the Fontan operation was 2.3 years, 25th and 75th percentiles 1.9 years and 2.9 years, respectively. The male-to-female ratio was 1.9:1.

Cardiopulmonary bypass was used for the various HLHS procedures. As a measure of data validity, only 5.8% (20 of 348 patients) of HLHS records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 64.9% (213 of 328 patients) used cross-clamping, whereas 0.9% (3 of 328 patients) used induced fibrillation; 34.2% (112 of 328) of the records were missing data on the specific technique used. Of the 213 patient records with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 46.9% (100 of 213 patients), crystalloid cardioplegia in 25.3% (54 of 213), and other cardioplegia in 7.5% (16 of 213 patients); data were missing in 20.2% (43 of 213) of the records. These data referable to CPB, aortic cross-clamping, and myocardial preservation are difficult to interpret because (1) donor ischemic time may not be the same as aortic cross-clamp time, (2) circulatory arrest time was not included in the analysis, and (3) there were a high number of missing entries. Because of these confounding issues, no attempt to subdivide the data by procedure was made.

Of the 348 records in the study period, 192 (55.1%) reported no complications. There were 61 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality rate of 17.5%. When subdivided by procedure, the operative mortality was 17.0% (10 of 59 patients) for cardiac transplantation, 27.3% (51 of 187) for the Norwood procedure, 0% (0 of 73) for bidirectional Glenn/Hemi-Fontan, and 0% (0 of 29 patients) for the Fontan operation. Of note, this database could not account for preoperative deaths attributable to either the transplant or Norwood groups. In addition, midterm deaths en route to the bidirectional Glenn or Fontan procedures could not be tracked. Of the 348 patients undergoing procedures, 37 of 59 (62.7%), 82 of 187 (43.9%), 54 of 73 (74.0%), and 19 of 29 patients (65.5%) were free of perioperative complications in the groups referable to cardiac transplantation, Norwood procedure, bidirectional Glenn/Hemi-Fontan, and Fontan procedure, respectively. About 31% (59 of 187) of Norwood patients underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. These data reflect the high incidence of intentional delayed sternal closure, which has become routine of late in some centers. In contrast, the reoperation rate for cardiac transplantation, bidirectional Glenn/Hemi-Fontan, and Fontan are 10.2% (6 of 59 patients), 12.3% (9 of 73), and 3.5% (1 of 29 patients), respectively. When considering the entire cohort of patients, complications attributable to infection occurred in 23 patients (6.6%). Other major system complications were reported in 4 (1.2%) neurologic complications, 58 (16.6%) pulmonary complications, and 7 (2%) renal complications. One hundred seven (30.7%) patients had complications other than those specifically mentioned.

Mean preoperative LOS for the transplant group was 22.9 days, 25th and 75th percentiles were 3 and 30 days, respectively. These data are in contrast to the mean preoperative LOS for the Norwood group, which was 5.6 days, 25th and 75th percentiles were 3 and 6 days, respectively.

Mean postoperative LOS reflected the severity of illness of the infant HLHS groups as opposed to those already palliated patients who presented for bidirectional Glenn/Hemi-Fontan or Fontan procedures. The mean LOS data for the transplant, Norwood, bidirectional Glenn/Hemi-Fontan, and Fontan groups were 30.2 days, 27.0 days, 9.4 days, and 12.4 days, respectively. The 25th and 75th percentiles for each group were 15 and 41 days, 16 and 30 days, 4 and 13 days, and 10 and 11 days, respectively.

This analysis of HLHS, which relates the two main treatment options, has revealed valuable data that are largely confirmatory of published reports [31, 32]. Tracking such clinical results can monitor improved outcomes that are being achieved at a greater number of institutions due to critical analysis and improved techniques.

Single ventricle/tricuspid atresia
Parameters for inclusion in the single ventricle/tricuspid atresia (SV-TA) study population were primary diagnosis of SV or TA with RA/caval to PA connection (Fontan/Kreutzer). Allowable concomitant diagnoses included ASD, VSD, PDA, AVC, TAPVC, PAPVC, all anomalies of systemic venous connection, RV to pulmonary outflow tract obstruction-vessel, aortic stenosis (any type: valvular, supravalvular, subvalvular, left ventricular outflow tract obstruction), CoA, levocardia, mesocardia, dextrocardia, situs solitus, situs inversus, atrial reentry tachycardia, Wolff-Parkinson-White syndrome, and Ebstein’s anomaly. No effort was made to quantify or identify the operations leading up to the Fontan operation, which was the focal point of this analysis. As a result, important risk factors, such as timing of PAB, status of the pulmonary arteries, extent of valvular dysfunction, and degree of ventricular dysfunction were not considered.

One hundred forty-seven patients emerged from the inclusion criteria. We arbitrarily decided to group them into (1) SV types, which included all types of effectively single right and left ventricles and (2) TA types, which were segregated from the rest of the single left ventricle types. In the 4-year study (1994–1997), 147 patients were reported with SV-TA. Ninety-two patients (62.6%) had SV and 55 patients (37.4%) had TA. The greater majority of SV-TA patients, 106 of 147 (72.1%) underwent diagnosis by both cardiac catheterization and echocardiography. Echocardiography only was used in 25 of 147 patients (17.0%), cardiac catheterization only was used in 9 of 147 (6.1%), and 7 of 147 patients (4.8%) had missing methods. These trends were consistent over the 4-year study period (1994–1997) and there was no appreciable difference between the SV and TA groups. The median age at operation (Fontan repair) for the SV-TA cohort was 3.3 years, 25th and 75th percentiles 2.1 and 5.7 years, respectively. Interestingly, the median ages at operation for SV and TA were similar at 3.5 years and 3.1 years, respectively. The female-to-male ratio was .5:1 for the SV group and 1.1:1 for the TA group.

Cardiopulmonary bypass was used for repair of SV-TA. As a measure of data validity, 2.0% (3 of 147 patients) of SV-TA records had missing CPB data (time of CPB run). Further analysis of those records with CPB data revealed 70.8% (102 of 144 patients) used aortic cross-clamping, whereas 5.6% (8 of 144 patients) used induced fibrillation; 23.6% (34 of 144) of the records were missing data on the specific technique used.

Of the 102 patients with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 50.0% (51 patients), crystalloid cardioplegia in 21.6% (22), and other forms of cardioplegia in 20.6% (21 patients); data were missing in 7.8% (8 patients) of the records.

The database input form queried Fontan patients for the use of fenestrations and lateral tunnels. No provision was made for the extracardiac Fontan, which is gaining in popularity at present. It also should be noted that there was a choice for "fenestration" and "lateral tunnel." Therefore, when fenestration is not chosen, we assume that a fenestration was not used or that it was in fact used but the input process failed to note it. This kind of assumption may in fact underreport the use of fenestration. The same is true for the lateral tunnel entry, which has the same disclaimer, as does the fenestration option. In addition, the way that the analysis was performed, we do not know which patients had both fenestration and lateral tunnel, although a sizeable number had to have both. Of the 147 patients in the SV-TA group, 37 (25.2%) had fenestrations and 110 (74.8%) did not have fenestrations or represent missing entries. Of these same 147 patients in the SV-TA group, 21.8% (32 of 147) had a lateral tunnel construction and 78.2% (115 of 147) did not have a lateral tunnel or represent missing entries. One might expect a lesser use of fenestration and "lateral tunnel" in the TA group. This turned out not to be the case, as fenestration was used in 25.0% and 25.5% of patients for the SV and TA groups, respectively. Likewise, the "lateral tunnel" reconstruction was used in 22.8% and 20.0% of patients for the SV and TA groups, respectively.

The complication incidence in SV-TA mirrored the more recently published reports [33]. Of the 147 records in the study period, 108 (73.5%) reported no complications. This ranged from a high of 88.2% (45 of 51 patients) in 1994 to a low of 54.3% (19 of 35 patients) in 1996. There were 7 operative deaths (death within hospitalization or within 30 days of surgical procedure) for a mortality of 4.8% over the 4-year study period (1994–1997). Curiously, the mortality was higher in the TA group, 4 of 55 (7.3%) compared to the SV group, 3 of 92 (3.3%). Fourteen of 147 patients (9.5%) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Pulmonary complications were reported in 6 patients (4.1%). Five patients (3.4%) had complications attributable to infection. Other major system complications were reported in a small number of patients: 1 major neurologic complication (0.7%), and no renal complications. In addition to or other than those major system or operative complications specifically mentioned, 30 patients (20.4%) were marked as having "other" complications.

Postoperative LOS for the SV-TA groups was a mean of 10.3 days, 25th and 75th percentiles, 7 and 10 days, respectively. As one might expect the mean LOS for the SV cohort was 11.7 days, 25th and 75th percentiles, 8 and 14 days, respectively, whereas the mean LOS for the TA cohort was 6.7 days, 25th and 75th percentiles, 6 and 8 days, respectively.

The trends noted in this analysis were consistent with the available large surgical series [33, 34]. More specific risk factors, such as extent of pleural effusions, types of postoperative arrhythmias, and the use of extracardiac conduits, could not be discovered by the database. Future database schemes will require updated data entry choices to assess the newer modifications of the Fontan operation for complex forms of single ventricle.

Corrected transposition of the great arteries
Inclusionary criteria for biventricular repair of corrected transposition of the great arteries (L-TGA) were primary diagnosis of L-TGA with VSD or VSD/PS with two ventricles. Allowable concomitant diagnoses included ASD, VSD, PDA, PA-VSD, Ebstein’s anomaly, congenital heart block, atrial tachycardia, ventricular tachycardia, Wolff-Parkinson-White syndrome, and left superior vena cava. The purpose of this analysis, beside assessing the standard outcomes, was to sample the types of corrective pressures that were being performed for this challenging set of patients, in light of the newer avant garde operations, such as "double-switch" [35] or "Mustard-Rastelli" [36] procedures. No attempt was made to evaluate or enumerate the prerepair procedures. For the purpose of comparison, patients with L-TGA with two ventricles were segregated in those with VSD (L-TGA-VSD) (14 of 33, 42.4%) and those with VSD and PS (L-TGA-VSD-PS) (19 of 33, 57.1%) over the 4-year study period (1994–1997). The division makes sense because there are different operations for each category. Most patients, 22 of 33 (66.7%), underwent diagnostic cardiac catheterization and echocardiography, whereas 10 of 33 patients (30.3%) underwent only echocardiography; no patient had only cardiac catheterization for diagnostic evaluation and there was 1 of 33 patients (3.0%) with missing method.

Median age at repair for both groups (L-TGA-VSD and L-TGA-VSD-PS) was 2.3 years, 25th and 75th percentiles 1.0 and 5.8 years, respectively. Patients with L-TGA-VSD had operative repair earlier than the L-TGA-VSD-PS cohort; median ages of 1.8 years (25th and 75th percentiles 0.5 and 3.6 years, respectively) and 3.8 years (25th and 75th percentiles 1.6 years and 11.0 years, respectively). The male-to-female ratio was 1.8:1.

Cardiopulmonary bypass data (time of CPB run) was missing in 5 of 33 patients (15.2%) of patient records; 28 of 33 patients (84.8%) had recorded CPB entries. Of these 28 patient records, 25 of 28 (89.3%) had cardioplegia entries; 3 of 28 patients (10.7%) had missing entries for cardioplegia. The types of cardioplegia were evenly dispersed: 9 of 25 (36.0%) used blood cardioplegia; 5 of 25 (20.0%) used crystalloid cardioplegia; 9 of 25 (36.0%) used "other" cardioplegia strategies; 2 of 25 (8.0%) represented missing entries.

Patients with L-TGA-VSD and L-TGA-VSD-PS have been treated with a wide variety of surgical options including a Fontan repair despite the fact that these patients have two functioning ventricles. Over the 4-year study period 14 patients with L-TGA-VSD underwent VSD closure in 4 of 14 patients (28.6%), Mustard/Senning-Rastelli procedure in 0 of 14 (0.0%), Mustard/Senning-Switch procedure in 1 of 14 (7.1%), Fontan procedure in 2 of 14 (14.3%), and 7 of 14 (50%) of patient records reported "other repairs." We did not isolate and interrogate these "other repairs" patient records. It is left to speculation whether there are indeed so many "other repairs" that were not sampled or that there are intrinsic flaws in the database analysis that did not "find" what these "other" patients had for therapeutic interventions. The results were worse for the L-TGA-VSD-PS group. Over the 4-year study period (1994–1997) 19 patients with L-TGA-VSD-PS underwent VSD closure and pulmonary valvotomy in none (0%), VSD closure and LV-PA conduit in 5 (26.3%). Mustard/Senning-Rastelli in none (0%), Fontan operations in 2 (10.5%), and a disappointing 12 (63.2%) had "other repairs," which of course did not fit into our programmed therapeutic choices. Likewise, we did not isolate and interrogate these "other repairs" patient records, leaving to speculation as to the merits/faults of the data-gathering system, data analysis program, and human omission/commission errors.

Operative mortality (death within hospitalization or within 30 days of surgical procedure) was low for this difficult group of patients (1 patient in each group; 2 of 33, 6.1%). Slightly more than 9% (3 of 33) of patients underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. No patient record listed complications due to infection, neurologic, pulmonary, or renal problems. Fully 75.8% (25 of 33 patients) reported no complications and 6 of 33 patients noted other complications than those listed. The postoperative LOS was similar in both groups and together was a mean of 6.2 days, 25th and 75th percentiles 4 and 8 days, respectively.

This analysis was disappointing because of the large number of "other repairs" that we did not capture. Future database schemes will have to be organized in such a manner as to discriminate between the large number of therapeutic options for these very interesting patients.

Double outlet right ventricle
Parameters for inclusion in the double outlet right ventricle (DORV) subgroup study population were primary diagnosis of DORV with two ventricles with the indicated subtype. Allowable concomitant diagnoses included ASD, VSD, PDA, TOF, transposition of the great arteries, RV-pulmonary outflow tract obstruction, and all anomalies of systemic venous connection.

The distribution of DORV in the STS study population reflects that described by other researchers [37, 38]. In the 4-year study period (1994–1997), 151 patients were reported, 33.1% (50 patients) with subpulmonic VSD, 44.3% (67) with subaortic VSD, 11.3% (17) with doubly committed VSD, and 11.3% (17 patients) with uncommitted VSD. The database entry system did not allow inclusion or exclusion of pulmonary stenosis in the primary diagnosis; likewise, there were no discriminating factors in the primary diagnosis for malposed or normally related great arteries. The majority of patients, 92 of 151 (60.9%), which spanned all VSD subtypes, underwent diagnosis by both cardiac catheterization and echocardiography. The remaining patient records had diagnosis by echocardiography only in 47 of 151 patients (31.1%) and by cardiac catheterization only in 8 of 151 patients (5.3%); 4 of 151 patients (2.7%) had missing methods.

Median age at repair of all types of DORV was 0.69 years (a little more than 8 months); 25th and 75th percentiles were 0.23 and 1.8 years, respectively. The subpulmonic VSD subtype was operated on at the earliest age; median age of repair was 0.39 years with 25th and 75th percentiles 0.06 (about 3.5 months) and 1.3 years, respectively. Median ages of operation for the subaortic, uncommitted, and doubly committed VSD types were 0.66, 0.83, and 1.31 years, respectively. The male-to-female ratio for the entire group was 1.3:1.

Cardiopulmonary bypass was used for repair of DORV. As a measure of data validity, 19.9% (30 of 151) of DORV records had missing CPB data (time of CPB run). Further analysis of these records with CPB data revealed 87.6% (106 of 121 patients) used cross-clamping of the aorta; 12.4% (15 of 121 patients) of the records were missing data on the specific technique used.

Of the 106 patients with cross-clamp times specified for the 4-year study period, blood cardioplegia was used in 64.1% (68 of 106 patients), crystalloid cardioplegia in 20.8% (22 of 106), and other cardioplegia in 9.4% (10 of 106 patients); data were missing in 5.7% (6 of 106 patients) of the records.

The therapeutic choices for repair of DORV with two ventricles are many. We sampled the expected modern-day operations that would include arterial switch, 15 of 155 patients (some patients in this category had more than one operation; 9.9%); Rastelli procedure, 13 of 155 (8.6%); VSD closure, PS repair with transannular patch, 2 of 155 (1.3%); VSD closure, PS repair without transannular patch, 16 of 155 (10.6%); Kawashima repair, 3 of 155 (2.0%); Fontan procedure, 7 of 155 (4.6%); and other procedures, 99 of 155 patients (65.6%). The "other procedures" category is extremely high and clearly represents a significant flaw in this analysis. Retrospectively, we discovered that we did not include VSD closure as a separate therapeutic modality that may have accounted for many of the "other procedures." This not withstanding, one is led to believe that errors defining inclusion criteria are responsible for these large omissions. Future database analyses must take into consideration the complexity of DORV and establish inclusion criteria that will maximize data reliance.

The complication incidence in DORV repair is surprisingly low. For instance, the overall mortality (death within hospitalization or within 30 days of surgical procedure) for the 4-year study period was 2.0% (3 of 151 patients). No complications were reported in 68.2% (103 of 151 patients), and 7.3% (11 of 151 patients) underwent reoperation for bleeding, valve dysfunction, residual defect, other cardiac or noncardiac problems, and/or sternal closure. Complications due to infection, neurologic events, pulmonary problems, and renal insufficiency occurred in 2.7% (4 of 151 patients), 0.7% (1 of 151), 11.3% (17 of 151), and 0.7% (1 of 151 patients), respectively. Twenty-nine of 151 patients (19.2%) reported "other complications" not specifically mentioned.

Mean postoperative LOS for all the VSD subtypes over the 4-year study period (1994–1997) was 12.5 days, 25th and 75th percentiles 6 and 13 days, respectively. The mean postoperative LOS for the VSD subtypes were 15.6 days for patients with subpulmonic VSD, 11.2 days for patients with subaortic VSD, 11.3 days for patients with doubly committed VSD, and 8.9 days for patients with uncommitted VSD.

This data analysis was disappointing because of the high number of "other procedures," which did not allow outcome evaluation of the different subgroups. The low mortality, although excellent in the reporting, seems too low for such a complex set of patients. Future database projects must address these problems for a proper evaluation of DORV.

	Summary

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
The cogent perioperative statistics from the STS National Congenital Heart Surgery—Summit Medical Systems Database Report (1994–1997) were analyzed for the 18 congenital heart categories. Statistical analyses were not performed due to the high number of "missing data points" (as high as 10%) that can confound data analysis and skew conclusions. We chose to provide the actual figures and percentages relating to each category. This not withstanding, several notable benefits, trends, and drawbacks can be identified.

The clear benefits of such a database are (1) that enough discriminating data can be collected to perform risk stratification, (2) that much of the data make sense and are largely confirmatory of major reports from individual institutions, and (3) that this database serves as a successful first stage step to an on-line, cyberspace-based international congenital heart surgery database that can fulfill the multiple needs of our profession and society.

	Congenital database subcommittee members

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
Constantine Mavroudis, MD, Chair, Congenital Database Subcommittee, Children’s Memorial Hospital, Chicago, IL; Carl L. Backer, MD, Children’s Memorial Hospital, Chicago, IL; Edward L. Bove, MD, University of Michigan Hospitals, Ann Arbor, MI; Duke E. Cameron, MD, The Johns Hopkins Hospital, Baltimore, MD; Fred H. Edwards, MD, University of Florida, Jacksonville, FL; Frederick L. Grover, MD, Ex Officio, University of Health Services Center, Denver, CO; John W. Hammon, Jr, MD, Bowman-Gray Medical Center, Winston-Salem, NC; John E. Mayer, MD, Children’s Hospital Medical Center, Boston, MA; W. Steves Ring, MD, University of Texas Southwestern Medical Center, Dallas, TX; Donald C. Watson, Jr, MD, Heart Center, Memphis, TN.

	Acknowledgments

 
We thank the following Congenital Heart Surgery Database Subcommittee members for their contribution to this project: Drs Carl L. Backer, Edward L. Bove, Duke E. Cameron, Fred H. Edwards, Frederick L. Grover, John W. Hammon Jr, John E. Mayer, and Donald C. Watson Jr., Cathy Sun, MS, provided valuable bio-statistical analysis for this project.

	Appendix 1. Names of centers

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 
John E. Mayer Jr, MD, Childrens Hospital Medical Center, Boston, MA;

Richard A. Hopkins, MD, Miriam Hospital, Providence, RI;

Charles W. Benoit, MD, Pennsylvania State Geisinger Health System, Danville, PA;

John A. Spratt, MD, Medical College of Virginia, Richmond, VA;

Benson R. Wilcox, MD, University of North Carolina Hospitals, Chapel Hill, NC;

John M. Kratz, MD, Medical University of South Carolina, Charleston, SC;

Richard A. Perryman, MD, Cardiothoracic Surgery, Miami, FL;

William Walker, MD, University of Tennessee Medical Group, Memphis, TN;

Erle Austin, MD, Norton Hospital of Alliant Health System, Louisville, KY;

Dennis S. Durzinsky, MD, Medical College of Ohio, Toledo, OH;

Alexander S. Geha, MD, University Hospitals of Cleveland, Cleveland, OH;

Children’s Hospital of Michigan, Detroit, Michigan

Luis A. Tomatis, MD, Butterworth Hospital, Grand Rapids, MI;

Constantine Mavroudis, MD, Childrens Memorial Hospital, Chicago, IL;

Francisco J. Puga, MD, Mayo Clinic, Rochester, MN;

Robert Tate, MD, Cardiothoracic & Vascular Surgeons, Austin, TX;

Alexander Duff James, MD, Driscoll Childrens Hospital, Corpus Christi, TX;

Shreekanth V. Karwande, MD, University of Utah Medical Center, Salt Lake City, UT;

John N. Goodwin, MD, Children’s Hospital of Los Angeles, Los Angeles, CA;

Kevin Turley, MD, California Pacific Medical Center, San Francisco, CA;

Leonard L. Bailey, MD, Loma Linda University Medical Center, Loma Linda, CA;

L. Ennix Coyness, MD, Center for Cardiac Surgery, Berkeley, CA;

Frank L. Hanley, MD, University of California, San Francisco, San Francisco, CA;

Paul C. Cartier, MD, Laval Hospital, Quebec, Canada;

W. Steves Ring, MD, University of Texas Southwestern Medical School, Dallas, TX.

	Appendix 2. Database overview (1994–1997) and incidence of primary diagnoses

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

 

	References

Top Abstract Introduction Analysis of diagnoses Summary Congenital database subcommittee... Appendix 1. Names of... Appendix 2. Database overview... References

1. Society of Thoracic Surgeons and Summit Medical Systems. Data Analyses of The Society of Thoracic Surgeons National Congenital Cardiac Surgery Database 1994–1997. Minnetonka, MN: Summit Medical Systems, 1998.
2. Craig R.J., Selzer A. Natural history and prognosis of atrial septal defect. Circulation 1986;37:805-815.[Abstract/Free Full Text]
3. Hoffman J.I.E., Rudolph A.M. The natural history of ventricular septal defects in infancy. Am J Cardiol 1965;16:634-653.[Medline]
4. McNicholas K.W., Bowman F.O., Jr, Hayes C.J., et al. Surgical management of ventricular septal defects in infants. J Thorac Cardiovasc Surg 1978;75:346-353.[Abstract]
5. Backer C.L., Idriss F.S., Zales V.R., et al. Surgical management of the conal (supracristal) ventricular septal defect. J Thorac Cardiovasc Surg 1991;102:288-295.[Abstract]
6. Idriss F.S., Muster A.J., Paul M.H., Backer C.L., Mavroudis C. Ventricular septal defect with tricuspid pouch with and without transposition. J Thorac Cardiovasc Surg 1992;103:52-59.[Abstract]
7. Mavroudis C., Backer C.L. The two-patch technique for complete atrioventricular canal. Semin Thorac Cardiovasc Surg 1997;9:35-43.[Medline]
8. Permut L.C., Mehta V. Late results and reoperation after repair of complete and partial atrioventricular canal defect. Semin Thorac Cardiovasc Surg 1997;9:44-54.[Medline]
9. Van Praagh R., Van Praagh S. The anatomy of common aorticopulmonary trunk (truncus arteriosus communis) and its embryologic implications. Am J Cardiol 1965;16:406-425.[Medline]
10. Bove E.L., Lupinetti F.M., Pridjian A.K., et al. Results of a policy of primary repair of truncus arteriosus in the neonate. J Thorac Cardiovasc Surg 1993;105:1057-1065.[Abstract]
11. Hanley F.L., Heinemann M.K., Jonas R.A., et al. Repair of truncus arteriosus in the neonate. J Thorac Cardiovasc Surg 1993;105:1047-1056.[Abstract]
12. Shore D.F., Ho S.Y., Anderson R.H., et al. Aortopulmonary septal defect coexisting with ventricular septal defect and pulmonary arteria. Ann Thorac Surg 1983;35:132-137.[Abstract]
13. Backer C.L., Stout M.J., Zales V.R. Anomalous origin of the left coronary artery. J Thorac Cardiovasc Surg 1992;103:1049-1058.[Abstract]
14. Del Nido P.J., Duncan B.W., Mayer J.E., Jr, Wessel D.L., LaPierre R.A., Jonas R.A. Left ventricular assist device improves survival in children with left ventricular dysfunction after repair of anomalous origin of the left coronary artery from the pulmonary artery. Ann Thorac Surg 1999;67:169-172.[Abstract/Free Full Text]
15. Smith A., Arnold R., Anderson R.H., Wilkinson J.L., Qureshi S.A., McKay R. Anomalous origin of the left coronary artery from the pulmonary trunk. J Thorac Cardiovasc Surg 1989;98:16-24.[Abstract]
16. Backer C.L., Mavroudis C., Zias E.A., Amin Z., Weigel T.J. Repair of coarctation with resection and extended end-to-end anastomosis. Ann Thorac Surg 1998;66:1365-1370.[Abstract/Free Full Text]
17. Van Son J.A.M., Falk V., Schneider P., Smedts F., Mohr F.W. Repair of coarctation of the aorta in neonates and young infants. J Cardiol Surg 1997;12:139-146.
18. Prieto L.R., DeCamillo D.M., Konrad D.J., Scalet-Longworth L., Latson L.A. Comparison of cost and clinical outcome between transcatheter coil occlusion and surgical closure of isolated patent ductus arteriosus. Pediatrics 1998;101:1020-1024.[Abstract/Free Full Text]
19. Mavroudis C., Backer C.L., Gevitz M. Forty-six years of patent ductus arteriosus division at Children’s Memorial Hospital of Chicago—standards for comparison. Ann Surg 1994;220:402-410.[Medline]
20. Hanley F.L., Sade R.M., Blackstone E.H., Kirklin J.W., Freedom R.M., Nanda N.C. Recommendations from a multicentered study regarding morphology specific management of neonatal pulmonary atresia with intact ventricular septum. J Thorac Cardiovasc Surg 1993;105:406-423.[Abstract]
21. Puga F.J. Surgical treatment of pulmonary atresia with ventricular septal defect. Prog Ped Cardiol 1992;1:37.
22. McElhinney D.B., Reddy V.M., Hanley F.L. Tetralogy of Fallot with major aortopulmonary collaterals. Pediatr Cardiol 1998;19:289-296.[Medline]
23. Kirklin J.W., Blackstone E.H., Jonas R.A., et al. Morphologic and surgical determinants of outcome events after repair of tetralogy of Fallot and pulmonary stenosis. A two-institution study. J Thorac Cardiovasc Surg 1992;103:706-723.[Abstract]
24. Watterson K.G., Malm T.K., Karl T.R., Mee R.B. Absent pulmonary valve syndrome. Ann Thorac Surg 1992;54:1116-1119.[Abstract]
25. Bando K., Turrentine M.W., Ensing G.J., et al. Surgical management of total anomalous pulmonary venous connection. Thirty-year trends. Circulation 1996;94:12-16.
26. Lupinetti F.M., Kulik T.J., Beekman R.H., III, Crowley D.C., Bove E.L. Correction of total anomalous pulmonary venous connection in infancy. J Thorac Cardiovasc Surg 1993;106:880-885.[Abstract]
27. Mavroudis C., Backer C.L., Idriss F.S. Arterial switch for transposition of the great arteries and associated malposition anomalies (Review). Adv Cardiac Surg 1992;3:205-247.
28. Wernovsky G., Wypij D., Jonas R.A., et al. Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants. A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation 1995;92:2226-2235.[Abstract/Free Full Text]
29. Elkins R.C., Knott-Craig C.J., Ward K.E., Lane M. The Ross operation in children. Ann Thorac Surg 1998;65:496-502.[Abstract/Free Full Text]
30. Wang A., Harrison J.K., Bashore T.M. Balloon aortic valvuloplasty. Prog Cardiovasc Dis 1997;40:27-36.[Medline]
31. Bove E.L. Current status of staged reconstruction for hypoplastic left heart syndrome. Pediatr Cardiol 1998;19:308-315.[Medline]
32. Jacobs M.L., Rychik J., Murphy J.D., Nicolson S.C., Steven J.M., Norwood W.I. Results of Norwood’s operation for lesions other than hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 1995;110:1555-1561.[Abstract/Free Full Text]
33. Mayer J.E., Jr, Bridges N.D., Lock J.E., Hanley F.L., Jonas R.A., Castaneda A.R. Factors associated with marked reduction in mortality for Fontan operations in patients with single ventricle. J Thorac Cardiovasc Surg 1992;103:444-451.[Abstract]
34. Mavroudis C., Zales V.R., Backer C.L., Muster A.J., Latson L.A. Fenestrated Fontan with delayed catheter closure. Circulation 1992;86(suppl 2):85-92.
35. Helvind M.H., McCarthy J.F., Imamura M., et al. Ventriculo-arterial discordance. Eur J Cardiothorac Surg 1998;14:173-178.[Abstract/Free Full Text]
36. Ilbawi M.N., DeLeon W.Y., Backer C.L., et al. An alternative approach to the surgical management of physiologically corrected transposition with ventricular septal defect and pulmonary stenosis or atresia. J Thorac Cardiovasc Surg 1990;100:410-415.[Abstract]
37. Belli E., Serraf A., Lacour-Gayet F., et al. Biventricular repair for double-outlet right ventricle. Results and long-term follow-up. Circulation 1998;98:360-365.[Abstract/Free Full Text]
38. Kleinert S., Sano T., Weintraub R.G., Mee R.B., Karl T.R., Wilkinson J.L. Anatomic features and surgical strategies in double-outlet right ventricle. Circulation 1997;96:1233-1239.[Abstract/Free Full Text]

This article has been cited by other articles:

				L. Dos, A. Dadashev, D. Tanous, I. J. Ferreira-Gonzalez, K. Haberer, S. C. Siu, G. S. Van Arsdell, E. N. Oechslin, W. G. Williams, and C. K. Silversides Pulmonary valve replacement in repaired tetralogy of Fallot: Determinants of early postoperative adverse outcomes J. Thorac. Cardiovasc. Surg., September 1, 2009; 138(3): 553 - 559. [Abstract] [Full Text] [PDF]

				J. P. Jacobs, M. L. Jacobs, B. Maruszewski, F. G. Lacour-Gayet, D. R. Clarke, C. I. Tchervenkov, J. W. Gaynor, T. L. Spray, G. Stellin, M. J. Elliott, et al. Current Status of the European Association for Cardio-Thoracic Surgery and The Society of Thoracic Surgeons Congenital Heart Surgery Database Ann. Thorac. Surg., December 1, 2005; 80(6): 2278 - 2284. [Abstract] [Full Text] [PDF]

				J. P. Jacobs, F. G. Lacour-Gayet, M. L. Jacobs, D. R. Clarke, C. I. Tchervenkov, J. W. Gaynor, T. L. Spray, B. Maruszewski, G. Stellin, J. Gould, et al. Initial Application in The STS Congenital Database of Complexity Adjustment to Evaluate Surgical Case Mix and Results Ann. Thorac. Surg., May 1, 2005; 79(5): 1635 - 1649. [Abstract] [Full Text] [PDF]

				D. M. Shahian, E. H. Blackstone, F. H. Edwards, F. L. Grover, G. L. Grunkemeier, D. C. Naftel, S. A.M. Nashef, W. C. Nugent, and E. D. Peterson Cardiac Surgery Risk Models: A Position Article Ann. Thorac. Surg., November 1, 2004; 78(5): 1868 - 1877. [Abstract] [Full Text] [PDF]

				J. P. Jacobs, C. Mavroudis, M. L. Jacobs, F. G. Lacour-Gayet, C. I. Tchervenkov, J. W. Gaynor, D. R. Clarke, T. L. Spray, B. Maruszewski, G. Stellin, et al. Lessons learned from the data analysis of the second harvest (1998-2001) of the Society of Thoracic Surgeons (STS) Congenital Heart Surgery Database Eur. J. Cardiothorac. Surg., July 1, 2004; 26(1): 18 - 37. [Abstract] [Full Text] [PDF]

				J. S. Tweddell, G. M. Hoffman, K. A. Mussatto, R. T. Fedderly, S. Berger, R. D. B. Jaquiss, N. S. Ghanayem, S. J. Frisbee, and S. B. Litwin Improved Survival of Patients Undergoing Palliation of Hypoplastic Left Heart Syndrome: Lessons Learned From 115 Consecutive Patients Circulation, September 24, 2002; 106(12_suppl_1): I-82 - I-89. [Abstract] [Full Text] [PDF]

				J. W. Gaynor, J. P. Jacobs, M. L. Jacobs, M. J. Elliott, F. Lacour-Gayet, C. I. Tchervenkov, B. Maruszewski, and C. Mavroudis Congenital heart surgery nomenclature and database project: update and proposed data harvest Ann. Thorac. Surg., March 1, 2002; 73(3): 1016 - 1018. [Abstract] [Full Text] [PDF]

				B. Maruszewski, F. Lacour-Gayet, M. J. Elliott, J. W. Gaynor, J. P. Jacobs, M. L. Jacobs, C. I. Tchervenkov, H. Kurosawa, and C. Mavroudis Congenital heart surgery nomenclature and database project: update and proposed data harvest Eur. J. Cardiothorac. Surg., January 1, 2002; 21(1): 47 - 49. [Abstract] [Full Text] [PDF]

				J. F. Stark, S. Gallivan, K. Davis, J. R.L. Hamilton, J. L. Monro, J. C.S. Pollock, and K. G. Watterson Assessment of mortality rates for congenital heart defects and surgeons' performance Ann. Thorac. Surg., July 1, 2001; 72(1): 169 - 175. [Abstract] [Full Text] [PDF]

				C. Mavroudis and J. P. Jacobs Congenital Heart Surgery Nomenclature and Database Project: overview and minimum dataset Ann. Thorac. Surg., April 1, 2000; 69(4): S2 - 17. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Constantine Mavroudis Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mavroudis, C. Articles by Schwartz, M. Search for Related Content PubMed PubMed Citation Articles by Mavroudis, C. Articles by Schwartz, M.