The Society of Thoracic Surgeons Congenital Heart Surgery Database: 2019 Update on Research

Abstract

As the largest congenital and pediatric cardiac surgical clinical data registry in the world, the Society of Thoracic Surgeons Congenital Heart Surgery Database (STS CHSD) serves as a platform for reporting of outcomes and for quality improvement. In addition, it is an important source of data for clinical research and for innovations related to quality measurement. Each year, several teams of investigators undertake analyses of data in the STS CHSD pertaining to the surgical management of specific diagnostic and procedural groups, as well as specific processes of care and their associations with patient characteristics across centers participating in the STS CHSD, and with outcomes. Additional ongoing projects involve the development of new or refined metrics for quality measurement and reporting of outcomes and center level performance. The STS, through its Workforce for National Databases and the STS Research Center and Workforce on Research Development provides multiple pathways through which investigators may propose and carry out outcomes research projects based on STS CHSD data. This article reviews research published within the past year.

Introduction

More than 95% of centers in North America with programs for surgical management of pediatric and congenital heart disease (CHD) participate in the Society of Thoracic Surgeons Congenital Heart Surgery Database (STS CHSD). Between 2002 and 2018, data pertaining to a total of 515,659 operations have been entered. Most recently,158,887 operations were entered during the 4-year data collection cycle of January 1, 2015 through December 31, 2018. [1]. Since 2010, the STS CHSD has included an anesthesia module that was developed in conjunction with the Congenital Cardiac Anesthesia Society (CCAS).

The STS CHSD data collection platform is thoroughly reviewed, assessed, and updated periodically to ensure that data collection is relevant and up to date with respect to progress and innovations in the practice of surgery for CHD and is current with respect to contemporary perspectives in the broader context of outcomes reporting and quality measurement. Version 3.3 of the STS CHSD was used from January 2016 – December 2018 [2]. In January 2019, Version 3.41 of the STS CHSD was implemented [3]. These periodic updates are engineered in such a fashion that the utility of the “legacy” data previously entered into the database using earlier versions is maintained. As a result, investigators are able to carry out research that is relevant with respect to the most contemporary challenges and trends in patient management, at the same time being informative with respect to the evolution of practice and trends in outcomes across eras.

The Duke Clinical Research Institute (DCRI) serves as the data warehouse and analytic center for the STS CHSD. In addition to performing the biannual data analyses of all STS CHSD data collected over each preceding 4-year period and generating Feedback Reports to all participant centers twice per year, DCRI also provides statistical and analytic expertise for collaboration with the STS Workforce on National Databases and the STS Quality Measurement Task Force (QMTF) in the development of robust tools for quality measurement. A unique aspect of this collaboration is that it makes it possible to re-estimate coefficients of metrics such as the STS CHSD Mortality Risk Model twice yearly, based on data from the most current 4-year analytic window, to coincide with the production of each STS-CHSD participant Feedback Report [4,5]. Mechanisms such as this are key to the understanding and equitable reporting of outcomes data. Meanwhile, DCRI also serves as the analytic center for much of the research that is undertaken based on STS CHSD data, as described below.

STS Congenital Heart Surgery Database Research Updates

Research undertaken using the STS CHSD is broadly divided into two major categories: outcomes research, and quality measurement. Most STS CHSD-related outcome studies are initiated by database participants and their colleagues through submission of a proposal to the Congenital Heart Surgery Committee of the STS Access and Publications (A&P) Task Force, which reviews investigator-initiated proposals twice a year [6]. DCRI serves as the analytic center for proposals approved by the A&P Task Force, and members of the Congenital Heart Surgery Committee of the A&P Task Force are available to provide support to investigators. STS CHSD-based outcomes research can also be proposed through the STS Task Force on Funded Research (TFR), formerly the Task Force on Longitudinal Follow-Up and Linked Registries (TFLFLR), or the STS Participant User File (PUF) Research Program. The TFR pathway is appropriate for evaluation of proposals that involve linkage of STS CHSD to other registries or sources of administrative data. These studies are usually funded by investigator institutions or other external funding sources. The STS PUF program makes it possible to perform analysis at investigators’ institutions of national-scale deidentified data from the database. Proposals to the PUF Research Program may be submitted at any time and are reviewed on a continuous basis.

The following is a summary of STS CHSD-based research articles published in peer-reviewed journals during the past year. Those listed first represent the work products of STS member investigators who submitted research proposals to either the Congenital Heart Surgery Committee of the STS A&P Task Force or the STS PUF Research Program. Brief synopses of these recently published articles are followed by additional reports of investigations carried out following review and approval by the STS TFR, as well as other studies involving the use of STS CHSD data.

Surgical Management and Outcomes of Ebstein Anomaly in Neonates and Infants

Ebstein anomaly is a rare congenital malformation with a broad spectrum of morphology and clinical presentation. Symptomatic neonates and infants are generally considered to be at highest risk. Management strategies depend largely upon individual center experience. The objectives of a multi-institutional study by Holst and associates were to delineate preoperative characteristics of neonates and infants undergoing surgery for Ebstein anomaly, characterize procedures and operations performed on these neonates and infants, and describe the associated morbidity and mortality to gain insights into how preoperative factors and operative management strategies may impact outcomes [7]. The study analyzed data pertaining to 255 neonates and 239 infants who underwent surgery at 95 STS CHSD centers between 2010 and 2016. Outcomes of interest included Operative Mortality and a composite mortality/major morbidity endpoint defined as Operative Mortality or the occurrence of 1 or more of 6 previously defined major complications [8].

Prenatal diagnosis was common in both neonates (71.0%) and infants (61.8%). Pulmonary atresia was present in one-fourth of cases (31.0% of neonates and 18.8% of infants), and Wolff-Parkinson-White syndrome was diagnosed in 3.2%. Among neonates, 62% required mechanical ventilation, 6.3% presented with shock that was unresolved at the time of surgery, and 6% required mechanical circulatory support.

Cases were classified on the basis of the primary procedure of the index operation. Among neonates, the most frequently performed primary procedure was Ebstein repair (n=101, 39.6%). STS CHSD Data Specifications Version 3.3: indicate that Ebstein repair “may include, among other techniques, repositioning of the tricuspid valve, plication of the atrialized right ventricle, or right reduction atrioplasty. Often associated ASD’s may be closed and arrhythmias addressed with surgical ablation procedures. These procedures should be entered as separate procedure codes.” [9]. Ebstein repair was followed in frequency by systemic-to-pulmonary shunt (n=52, 20.4%), and tricuspid valve closure (n=24, 9.4). Of the latter, the majority included systemic-to-pulmonary shunt, ASD creation or enlargement, and PDA closure, as concomitant procedures. Ebstein repair was the primary procedure performed on most neonates that had preoperative extracorporeal membrane oxygenation (ECMO) support (50%, n = 8 of 16), mechanical ventilation (56.7%, n = 89 of 157), or preoperative shock (53.7%, n = 22 of 41). Three neonates underwent cardiac transplantation subsequent to the index operation but during the same hospitalization. Among infants, the most frequently performed primary procedure was a superior cavopulmonary anastomosis (SCPA; n = 91, 38.1%), most often in association with concomitant shunt ligation and takedown (n = 55; 60.4% of SCPA). Ebstein repair was the second most frequently performed primary procedure among infants (n = 37, 15.5%). Other commonly performed primary procedures included pulmonary artery banding, systemic-to-pulmonary shunt, and repair of ventricular septal defect. Cardiac transplantation was the index operation in 6 infants.

Overall Operative Mortality was 18.6% (92 of 494 patients), with Operative Mortality of 27.5% (70 of 255) among neonates and 9.21% (22 of 239) among infants. The composite mortality/major morbidity endpoint was 51.4% (131 of 255) among neonates and 20.1% (48 of 239) among infants. Reoperation during the episode of care was 63.1% (161 of 255) among neonates. While the most frequently performed reoperation was delayed sternal closure (106 of 255; 41.6%), systemic-to-pulmonary artery shunt, or shunt revision was performed in 37 neonates (14.5%), and shunt combined with tricuspid valve closure was performed at reoperation in 8 neonates (3.1%). Among infants, the most frequently performed reoperations were delayed sternal closure (24 of 239, 10.0%). Systemic-to-pulmonary artery shunt, or shunt revision was performed at reoperation in 7 infants (2.9%).

Mortality for neonates requiring preoperative ECMO was 75.0% (12 of 16) versus 0% (0 of 2) for infants. For neonates with preoperative shock, mortality was 46.3% (19 of 41) versus 25.0% (3 of 12) for infants. Mortality for neonates with preoperative ventilation was 34.4% (54 of 157) versus 21.2% (7 of 33) for infants. When mortality was stratified by delayed sternal closure versus primary sternal closure, there was no difference among neonates (30.2% versus 25.5%; p = 0.47), while mortality was higher in infants with delayed sternal closure than with primary sternal closure (20.8% versus 7.9%; p = 0.054). Neither Operative Mortality nor the composite mortality/major morbidity endpoint was significantly different between groups with or without pulmonary atresia, for neonates, infants, or all patients.

The report by Holst and associates is the largest series of patients undergoing operative management of Ebstein anomaly in the first year of life and highlights the breadth of variation in both the preoperative status of these patients and the operative strategies used in their management. Neonates and infants presented quite differently, with neonates being much more likely to be on preoperative ECMO, have preoperative shock, and/or require preoperative mechanical ventilation. These differences likely reflect differences in underlying disease severity and were strongly associated with outcomes. Operative intervention in critically ill neonates, while risky, is often mandated by condition of the patient and is often utilized in the setting of failure of medical management. The rarity of Ebstein anomaly continues to largely limit surgical experience at most individual centers. This multi-institutional report provides new insights regarding the surgical landscape of neonates and infants with Ebstein anomaly. More detailed assessment of potential interdependence of anatomy, physiology, operative strategies, and late outcomes would best be addressed by a dedicated multi-center prospective study.

National Benchmarks for Proportions of Patients Receiving Blood Transfusions During Pediatric and Congenital Heart Surgery

Kartha and associates undertook an analysis of data pertaining to index cardiopulmonary bypass (CPB) operations in the STS CHSD (2014 through 2015) to determine national benchmarks and to assess variability across centers with regard to the proportion of patients receiving intraoperative transfusion of packed red blood cells (PRBC) [10]. The study encompassed 22,874 index CPB operations at 81 centers. Center-level intraoperative PRBC transfusion rates were stratified by age group, weight, STAT Mortality Category, and lowest core temperature. Among neonates, patients weighing less than 5 kg, STAT Category 5 operations, and operations with lowest core temperature below 20 degrees C, the median of values across centers for the proportion of cases involving intraoperative PRBC transfusion(s) approached 100%, with very narrow IQRs. Center PRBC transfusion rates declined with increasing patient age, but with greater variability (wider interquartile ranges) across centers. Intraoperative PRBC transfusion was uncommon (median center transfusion rates <30%) in older patients (teenagers and adults) undergoing lower complexity (STAT Mortality Category <3) operations.

Nationwide benchmarks regarding intraoperative blood transfusion practices in pediatric and congenital heart surgery had not previously been clearly established. While adherence to restrictive transfusion practices is becoming increasingly widespread in adult cardiac surgery, less is known about the effect of restrictive versus liberal transfusion protocols in children undergoing open heart surgery. Arguments in favor of restrictive transfusion practices must be balanced with data that raise concerns about lower hematocrit during CPB. In a randomized trial of infants undergoing cardiac surgery with hypothermic CPB, Jonas and colleagues [11] reported that infants assigned to the lower hematocrit group had worse perioperative outcomes and worse developmental outcomes. The analysis also revealed that use of a higher hematocrit level during hypothermic CPB was not associated with greater use of blood products. The STS multi-center analysis by Kartha and associates has revealed that in 2014–2015, on a national level, neonates and infants were likely to receive PRBCs during operations using CPB regardless of STAT Mortality Category. In toddlers (>1 year to <4 years), on the other hand, there was some variability in blood use, particularly in STAT Mortality Category 1 and 2. As patient age increased, rates of transfusion decreased and variability increased.

During the relevant time period of this study, CHSD data were collected in such a way that it was most practical to base the analysis on the binary assessment of whether or not a patient received intraoperative PRBC transfusion(s). The STS CHSD Data Collection Form has since been revised so that future analyses can focus on the actual amount of blood transfused to each patient, the varieties of types of products administered, and the phase of care during which each transfusion was initiated.

A Study of Practice Behavior for Endotracheal Intubation Site for Children with Congenital Heart Disease Undergoing Surgery

Greene and associates analyzed data from the STS CHSD with the objective of describing the practice of nasal intubation in the pediatric population undergoing cardiac surgery with CPB and assessing the risks/benefits of a nasal versus oral intubation [12]. The study population consisted of patients <18 years of age for whom comprehensive data regarding endotracheal intubation was entered in the Anesthesia Module of the STS CHSD for operations between January 2010 and December 2015. Patients with a preoperative endotracheal tube, tracheostomy, or known airway anomalies were excluded. Associations between route of tracheal intubation (nasal vs oral) and a composite measure of infection risk (wound infection, mediastinitis, septicemia, pneumonia, and endocarditis) was assessed using multivariable modeling with adjustment for important patient characteristics (e.g., weight, age, comorbidities), case complexity, and center effects. Other outcomes assessed included: length of intubation, hospital length of stay, and airway complications including accidental extubations.

Analysis revealed that nasal intubation was used in 41% of operations in neonates, 38% in infants, 15% in school-aged children, and 2% in adolescents. Nasal intubation appeared protective against accidental extubation in neonates (P = .02), but not in other age groups. Multivariable analysis in neonates and infants showed that the nasal route of intubation was not associated with the infection composite (relative risk [RR], 0.84; 95% CI, 0.59–1.18) or a shorter length of stay (RR, 0.992; 95% CI, 0.947–1.039), but was associated with a shorter intubation length (RR, 0.929; 95% CI, 0.869–0.992).

The study revealed the lack of a standardized approach to airway management in children undergoing heart surgery. Importantly, multivariable analysis showed no significant association between a nasal route of intubation and infection when neonates and infants were considered together as a population. Regarding other complications, neonates who received oral intubation were more likely to have accidental extubation or transesophageal echocardiography (TEE)-related extubation than those receiving nasal intubation (0.4% vs 0.0%; P value .0177). Investigators concluded that their findings were supportive of the current practice whereby oral intubation is most prevalent in the older age group. In neonates and infants, however, the nasal route of intubation was not associated with the infection composite. And in neonates, the nasal route may be preferable on the basis of decreased risk of accidental extubation and a shorter length of intubation.

The Prevalence and Impact of Congenital Diaphragmatic Hernia Among Patients Undergoing Surgery for Congenital Heart Disease.

Congenital diaphragmatic hernia (CDH) has a prevalence of 2.55 of 10,000 births and even in isolation, it is associated with significant morbidity and mortality. Various investigators have suggested that as many as 20% of children with CDH have a concomitant congenital cardiac anomaly. A recent report from the Congenital Diaphragmatic Hernia Study Group database reported prevalence of cardiac defects among patients enrolled in the registry to be 18% [13]. In this analysis, patients with CDH and cardiac anomalies had lower survival rates than CDH patients with no evidence of congenital heart disease. The “other side of the coin,” i.e. the prevalence and impact of congenital diaphragmatic hernia (CDH) on children undergoing cardiac surgery had not been explored in a multi-center investigation. Fraser and associates [14] queried the STS CHSD to identify pediatric patients undergoing cardiac surgery who also had CDH. Of 157,419 patients <18 years of age undergoing an index cardiovascular operation at 122 centers between January 1, 2010 and December 31, 2016, there were 426 patients with a concomitant diagnosis of CDH (coded as a noncardiac congenital anatomic abnormality in the STS CHSD). Among these patients with CDH there were 89 neonates (0.25% of all neonatal index operations), 217 infants (0.39%), and 120 children (0.18%). In neonates and infants, the frequency of concomitant CDH was highest in patients with tetralogy of Fallot (0.45%, n = 41), coarctation (0.39%, n = 27), and complete atrioventricular septal defects (0.31%, n = 19).

Cardiac surgical outcomes among patients with a concomitant diagnosis of CDH were evaluated separately for each age group. Among neonates, those undergoing low complexity cardiac procedures experienced Operative Mortality twice as high (11.8% vs 5.7%) as neonates without CDH (p = 0.252; likely a reflection of the relatively small numbers). For neonates undergoing STAT 3–5 cardiac procedures, Operative Mortality was significantly higher in CDH patients (34.4% vs 10.3%, p < 0.001). In both low and high complexity cases, neonates with CDH were more likely to experience major postoperative complications. Neonates with CDH were also more likely to undergo an unplanned reintervention after their index cardiac procedure compared to neonates without CDH. Among infants, Operative Mortality was significantly higher for the CDH group irrespective of STAT category. However, in children older than 1 year, mortality did not differ significantly between those with and without CDH undergoing either low or high complexity cardiac procedures.

Fraser and associates are the first to document multi-center cardiac surgical outcomes in patients with CDH, broken down by age group as well as by cardiac diagnosis and primary cardiac procedure. Such information will help inform therapeutic decisions and will also be an important resource for parental counseling.

Prenatal Diagnosis Influences Preoperative Status in Neonates with Congenital Heart Disease

Quartermain and associates designed a study to assess the potential impact of a prenatal diagnosis (PND) of congenital heart disease (CHD) on the presence of major pre-operative risk factors among neonates undergoing heart surgery [15]. They analyzed data on 12,899 neonates in 11 common diagnostic groups undergoing index cardiac operations from 2010 through 2014 at 112 centers. They investigated the potential association between PND and the composite endpoint of nine major pre-operative risk factors, using conditional logistic regression models stratified by center to adjust for baseline co-variates from the STS CHSD Mortality Risk Model [4,5].

The overall rate of PND was 49% (6316/12,899) over the 5-year study period, with an increase in PND rates from 45% in 2010 to 53% in 2014. PND rates varied across defect types, with the highest rates for single ventricle defects (76%) and the lowest detection rates for total anomalous pulmonary venous connection (13%). Major pre-operative risk factors were present in 34% (5003/12,899) of the overall study population. Patients with a prenatal diagnosis of CHD had lower rates of major pre-operative risk factors (30.3% % vs. 37.0%, p < 0.0001) and had lower or equivalent rates of each of the component risk factors (shock, mechanical circulatory support, mechanical ventilatory support, renal dysfunction, neurologic deficit, cardiopulmonary resuscitation, hepatic dysfunction, necrotizing enterocolitis, and coagulopathy). In multivariable analysis adjusting for potential confounders (sex, race, weight-for-age-and-sex z-score, chromosomal abnormalities/syndromes/non-cardiac congenital anatomic anomalies, prematurity, and STAT category), major pre-operative risk factors were less prevalent among neonates with a PND diagnosis compared to neonates without PND (adjusted OR 0.62, 95% CI 0.57–0.68, p < 0.0001). Results were similar when evaluated separately in three specific diagnostic categories with adequate numbers for subgroup analysis: coarctation of the aorta, hypoplastic left heart syndrome, and transposition of the great arteries with intact ventricular septum. A sensitivity analysis was undertaken to assess whether the findings may be confounded by a higher prevalence of prematurity, genetic abnormalities, and/or non-cardiac congenital anatomic abnormalities in patients with a PND of CHD. Results of this subgroup analysis were consistent with those of the primary analysis.

While some studies have demonstrated a benefit of PND on clinical outcomes, several other studies have not supported that finding. The study by Quartermain clarifies an issue currently obscured by conflicting literature and a lack of definitive evidence that PND translates to improved clinical benefits across a wide range of heart defects. The study design focused on the relative prevalence of important pre-operative risk factors rather than on surgical outcomes, effectively removing the effects of variability in center-specific surgical care and performance that can impact outcome measures. This study, the largest to date assessing the potential impact of a PND on pre-operative status in patients with CHD, demonstrated that among patients with critical forms of CHD, a PND was associated with lower rates of major pre-operative risk factors.

Readmission After Pediatric Cardiothoracic Surgery

Kogon and associates analyzed data from STS CHSD to determine the prevalence of hospital readmission after pediatric cardiac surgery, describe patient characteristics, and evaluate risk factors for readmission [16]. Between January 2014 and December 2016, 56,429 patients < 18 years of age, at 100 centers, were discharged home following cardiac surgery. The primary outcome variable was readmission within 30 days of discharge. Overall, 6,208 (11%) patients were readmitted.

Baseline characteristics, preoperative factors, perioperative data, and postoperative outcomes data were collected for the overall cohort. Regression analysis was used to identify factors associated with readmission, with adjustment for covariates in the STS CHSD Mortality Risk Model [4,5] and additional study-specific covariates including race, ethnicity, use of CPB, CPB time, postoperative length of stay (LOS), any major postoperative complication, prolonged postoperative ventilation, a composite of any other postoperative complication, and day of the week of discharge. Factors associated with increased odds of readmission included the presence of noncardiac congenital anatomic abnormalities (odds ratio [OR], 1.24), chromosomal abnormalities or genetic syndromes (OR, 1.24), preoperative mechanical circulatory support (OR, 1.36), other preoperative factors (OR, 1.21), prior cardiac surgery (OR, 1.31), Hispanic ethnicity (OR, 1.13), higher STAT Category (STAT level 3 vs 1, OR, 1.22; STAT 4 vs 1, OR, 1.48; STAT 5 vs 1, OR, 2.62), prolonged postoperative LOS (OR, 1.07 per day from 0 to 14 days; OR, 1.01 per week >14 days), any major complication (OR, 1.27), any other postoperative complications (OR, 2.00), and discharge on a weekday (OR, 1.07).

For the entire cohort, the most common primary reasons for readmission were respiratory or airway complications (15.3% of readmissions), septic or infectious complications (12.7%), and other readmissions not related to the surgical procedure for the current admission (21.5%). This held true for the majority of the individual benchmark operation groups as well, with the exception of the Glenn/hemi-Fontan and Fontan groups. Pleural effusions accounted for 23.7% of readmissions in the Glenn/hemi-Fontan group and 62.8% in the Fontan group. Information from this largest-to-date study of hospital readmissions following pediatric cardiac surgery may help support the development of targeted interventions to reduce the rate of occurrence and impact of some avoidable hospital readmissions after congenital heart surgery.

Postoperative Transcatheter Interventions in Children Undergoing Congenital Heart Surgery.

Postoperative transcatheter interventions (TCIs) are performed after congenital heart surgery to treat residual or recurrent anatomic lesions. While current guidelines indicate that it is reasonable to perform TCIs in patients that are not following the expected course, knowledge gaps exist with respect to the frequency with which these procedures are performed, center variation in the use of TCIs, and whether center variability might impact outcomes. To address these questions, Thibault et al analyzed data from the STS CHSD pertaining to 104,743 patients undergoing a cardiovascular surgical operation whose data were entered into the STSCHSD between 2010 and 2016 [17].

Performance of a postoperative TCI procedure was ascertained on the basis of the STS-CHSD complication variable “unplanned interventional therapeutic cardiovascular catheterization procedure during the postoperative or postprocedural time period” or on the basis of any procedure code for a therapeutic (interventional) catheterization at any time during the same hospitalization but following the index operation. A postoperative TCI was performed after 2.5% (n=2615) of all index operations, and 5.7% (n=1433) of all STAT Category 4 and 5 operations. Single ventricle patients accounted for 43% of all postoperative TCIs. Overall, the median age of patients undergoing TCI was 2.7 (IQR 0.2–8.0) months. Neonates accounted for 43% of all TCIs performed, and infants (<1 year, inclusive of neonates) accounted for 80%. Postoperative TCIs were performed after 5.2%, 2.5%, and 1.2% of all index operations performed in neonates (<30 days), infants (30 days to 1 year), and older children (ages, 1–18 years), respectively.

Variability in rates of performance of postoperative TCI across the 107 included centers was considerable, with risk-adjusted rates of postoperative TCI ranging from 0.0% to 8.0% overall and 0.0% to 20.7% for STAT 4 and 5 cases. Occurrence of postoperative TCI was associated with higher risk-adjusted odds of Operative Mortality (odds ratio, 4.06; 95% CI, 3.60–4.58). Centers with higher postoperative TCI rates had higher overall Operative Mortality (R²=0.23; p=0.02) but did not have higher post-TCI mortality (p=0.10). There was no correlation between center TCI rates and failure-to-rescue (p=0.19), defined as death following a major postoperative complication (as previously defined in the STS CHSD including acute renal failure requiring dialysis, neurological deficit persisting at discharge, atrioventricular block or arrhythmia requiring a permanent pacemaker, postoperative mechanical circulatory support, phrenic nerve injury, or any unplanned cardiac reintervention before discharge) [8]. For ascertainment of failure-to-rescue in this study, TCIs were not counted among unplanned cardiac reinterventions.

The study by Thibault and associates is the first multicenter study to focus on postoperative TCIs in children undergoing heart surgery. It reveals that patients undergoing postoperative TCI procedures have poor outcomes with a >4-fold increased risk-adjusted odds of mortality. The overall 20.2% Operative Mortality rate is consistent with prior single-center studies. These outcomes likely reflect the high risk associated with the residual or recurrent anatomic lesions that warrant postoperative TCI, more than procedural adverse events associated with TCI. Outcomes are similarly poor for patients undergoing unplanned cardiac operations in the postoperative period. In a recent analysis of the STS-CHSD, Costello et al reported 18.8% mortality for patients undergoing any cardiac reintervention (of which 70% were surgical reoperations) in the postoperative period [18]. Together, the findings of these two studies suggest that prevention and early recognition of residual/recurrent anatomic lesions are paramount to ensuring optimized outcomes.

Cardiac Surgery in Patients with Trisomy 13 and 18

While congenital heart disease is common in patients with Trisomy 13 (T 13; Edwards Syndrome) and Trisomy 18 (T 18; Patau syndrome), there has been a paucity of available data describing cardiac surgical outcomes in patients with these commonly occurring types of aneuploidy. Cooper and associates analyzed data from the STS CHSD pertaining to patients with T13 and T18 who underwent cardiac surgery (2010–2017) [19]. There were 343 operations (T13: n=73 and T18: n=270) performed on 304 patients. Among 125 hospitals submitting data, 87 (70%) performed at least 1 operation and 26 centers (30%) performed ≥5 operations on patients with T 13 or T 18. Nine centers (10.35) performed 10 or more operations.

The 73 operations in patients with T13 had a slight female majority (54%), with a prematurity rate of 18%. Overall median age was 4.5 months and median weight was 4.8 kg at time of operation. More than one-third (n=27, 37%) of the patients had a reported associated noncardiac anatomic abnormality. One-fifth (n=16, 21%) had undergone prior cardiothoracic surgery. Sixteen patients (22%) went to the operating room on mechanical ventilation. For T13 patients, hospital mortality was 11% (8 of 73 patients). The fraction of patients that experienced one or more postoperative complications was high at 58% (n=42) compared with 36% (n=72 345) for all STS CHSD patients, P<0.001.

The 270 operations in patients with T18 had a 3:1 female to male ratio, and a prematurity rate of 23.7%. Overall median age was 3.7 months and median weight was 3.5 kg at time of operation. Noncardiac abnormalities were common (n=125; 46.3%). Prior cardiac surgery operations had been performed in 16.3% (n=44) of patients. Nearly one-third (n=82, 30.4%) required mechanical ventilation before going to the operating room. For T18 patients, hospital mortality was 16% (42 of 270 patients). The postoperative complication rate was 55.8% (n=150). Preoperative mechanical ventilator support was associated with in-hospital mortality among patients with T13, (p=0.011) with an unadjusted odds ratio of 8.2. Likewise, it was associated with in-hospital mortality among. patients with T18, (p<0.0001) with an unadjusted odds ratio of 8.5.

The primary diagnoses among patients with T13 and T18 undergoing surgery were diverse, with ventricular septal defects being extremely common in patients with T18, representing over half of the primary diagnoses, and accounting for 29% of the primary diagnoses among patients with T13. While there were only 7 patients in STAT category 5, there were 91 in STAT category 4. That STAT categories 4 and 5 together represent 28.5% of all operations indicates a willingness to operate on patients with complex heart disease. Notably, there were 13 operations on patients with single ventricle disease including 2 Norwood operations, a bidirectional Glenn procedure, and 3 Fontan operations. While the information is limited to that pertaining to the surgical hospitalization, and does not address longer-term outcomes, the report by Cooper contains a great deal of well-organized descriptive data, and can serve as a useful resource for practitioners and centers contemplating management strategies for T13 and T18 patients with congenital heart disease.

Factors Associated with Adverse Outcomes after Repair of Anomalous Coronary from Pulmonary Artery

Straka and colleagues analyzed outcomes following surgical repair of Anomalous Coronary Artery from the Pulmonary Artery (ACAPA) in 703 patients between 2007–2016 utilizing the Participant User File (PUF) [20]. Study objectives were to compare patient characteristics, peri-operative factors, and outcomes between survivors and non-survivors of surgical repair of ACAPA using a multicenter dataset of ACAPA patients. Investigators also sought to determine the predictors of in-hospital mortality and need for post-operative ECMO following ACAPA repair.

There were no intra-operative deaths. Of 703 patients who underwent ACAPA repair, twenty (2.8%) died before hospital discharge, fifteen died within 30 days of surgery and an additional five within 95 days of surgery. Of these deaths, mortality rates among neonates (<30 days), infants (30 days to 1 year) and children (1–18 years) were 5.3%, 3.9% and 0.4%, respectively. The median number of days from surgery to mortality was 9 days (IQR: 4.5 – 27.5). Compared with survivors, the non-survivors were younger (median 2.4 vs. 4.8 months; p <0.001), weighed less (3.85 vs. 6.00 kg; p <0.001) and had lower body surface area (0.22 vs. 0.31 m2, p<0.001). There were no statistically significant differences in prematurity, race, ethnicity region, surgical era, or insurance coverage. The non-survivors were more likely to have had pre-operative mechanical ventilation (55.0% vs. 18.2%, p<0.001) and to have presented in shock (40.0% vs. 5.6%, p < 0.001) compared with survivors. Non-survivors had longer CPB times (198 vs. 125 minutes, p <0.001) and delayed sternal closure (50.0% vs. 28.3%, p=0.045) compared with survivors. The survivors were also noted to have 4 (0.6%) patients who underwent VAD implantation during or following ACAPA repair and 58 (8.5%) who underwent concomitant mitral valvuloplasty. In comparison, no patients in the non-survivor group underwent concomitant mitral valve repair, though these differences did not reach statistical significance. Compared with survivors, the non-survivors were more likely to have experienced bleeding requiring reoperation (50.0% vs. 3.5%, p<0.001), cardiac complications (55.0% vs. 20.5%; p<0.001), neurological complications (15.0% vs.0.9%; p=0.0015), multi-organ dysfunction syndrome (15.0% vs. 0%, p<0.001) or post-operative ECMO support (75.0% vs. 6.5%; p<0.001). The non-survivor group was also more likely to undergo planned (45.0% vs. 18.2%; p=0.006) and unplanned (25.0% vs. 1.6%; p<0.001) re-operation.

By multivariable analysis, weight <5 kg, longer CPB time, and presence of pre-operative shock were associated with in-hospital mortality. When the need for post-operative ECMO support was then introduced into the model to determine its relationship with in-hospital mortality, it appeared to be a strong marker of mortality (OR 11.8, 95% CI: 3.62, 38.4; p<0.001). With ECMO included in the model, presence of pre-operative shock (OR 4.55, 95% CI: 1.37, 15.1; p=0.01) and weight < 5 kg (OR 7.04; CI: 1.46, 33.9) remained statistically significant. Although longer CPB time was still associated with in-hospital mortality, it was no longer significant.

Analysis of risk factors associated with need for post-operative ECMO support identified lower weight, pre-operative shock, CPB time and delayed sternal closure as being significantly associated with the need for post-operative ECMO. Investigators suggested that the association of pre-operative shock and post-operative ECMO use with mortality is likely a marker of pre-operative hemodynamic instability related to ventricular dysfunction. They inferred that this association also points to the need for prompt surgical repair in ACAPA patients, and that it may suggest a role for the early deployment of mechanical circulatory support (ECMO or VAD) immediately following repair, particularly in younger patients presenting with significant hemodynamic compromise.

The Congenital Heart Technical Skill Study: Rationale and Design.

The Congenital Heart Technical Skill Study (CHTSS) was designed as a multi-center, multi-surgeon collaborative investigation to assess associations between peer ratings of direct, video observation of operating surgeons with contemporaneous surgeon-specific outcomes from a national clinical registry, the STS CHSD. A secondary aim was to assess potential associations between center and surgeon characteristics with patient outcomes, after adjusting for the effects of surgeon technical skill. The structured framework of an observational study allows the added benefit of providing operating surgeons with peer feedback on their technical skill. CHTSS investigators hypothesized that surgeon technical skill is one potential source of inter-surgeon variation in outcomes for children with congenital heart defects and that after accounting for differences in surgeon technical skill, one can better understand the effects of other provider- or institution-level factors

Because of planned future investigations relating to resource utilization, the study was initially opened to surgeons who were at least one year from terminal training and contributing data to The Society of Thoracic Surgeons–Congenital Heart Surgery Database (STS-CHSD) and The Pediatric Health Information System (PHIS) database. As of April 2018, it has been opened to all congenital heart surgeons contributing data to the STS-CHSD, whether or not they participate in the PHIS Database. Surgeons are the participants of this research. No patient identifiers are collected. The CHTSS proposal was reviewed and approved by the STS Task Force on Funded Research (TFR).

Anderson and associates published their first report on the progress of the CHTSS [21]. They described elements of study design, including video acquisition and review, which has been underway since April 2016. Each surgeon is asked to submit up to two representative videos of himself/herself (or a designated trainee) performing each of three operation types: (1) The Norwood/Damus-Kaye-Stansel (DKS) procedure for hypoplastic left heart syndrome, (2) atrioventricular canal repair for complete balanced atrioventricular canals (CAVCs), and (3) the arterial switch for dextro-transposition of the great arteries, with or without a ventricular septal defect. Videos are anonymously rated by peer surgeons using an instrument modified slightly from the Objective Structured Assessment of Technical Skills (OSATS) score [22] to reflect the skills of congenital heart surgeons. The CHTSS study evaluates the technical skills of attending pediatric cardiothoracic surgeons by direct, video observation, and a peer-review mechanism. By mid-2018, 30% of congenital heart surgeons operating in the United States from nearly half of eligible centers had agreed to participate in this study, suggesting a robust interest and enthusiasm in the congenital heart surgery community for critical assessment of current performance, as well as an eagerness to contribute to overall improvement in surgical performance.

Lessons learned in the use of clinical registry data in multi-center prospective studies: the Pediatric Heart Network Residual Lesion Score Study and the NIH-sponsored STRESS Trial.

The Residual Lesion Score Study is a prospective, multi-center, observational cohort study conducted by the Pediatric Heart Network to assess the association between residual lesions following specified cardiovascular surgical operations and early and midterm outcomes. The Residual Lesion Score (RLS) Study combined two methods for data collection: (1) the traditional method of data collection utilized by the Pediatric Heart Network, which is done by trained research staff and (2) the extraction of existing local registry data that are already being collected at the sites for the purpose of submission to the STS CHSD. This was the first prospective study within the Pediatric Heart Network to pilot the use of registry data for a portion of the study variables.

Prospero and colleagues reported specifically on the use of clinical registry data in this multi-center prospective study [23]. A survey addressing processes and perceptions was administered to RLS study site and data coordinating center staff. Survey response rate was 98% (54/55). Overall, 57% perceived that using registry data saved research staff time in the current study, and 74% perceived that it would save time in future studies; 55% noted significant upfront time in developing a methodology for extracting registry data. Investigators concluded that use of existing registry data was perceived to save time and promote efficiency. They noted that consideration must be given to the upfront investment of time and resources needed.

They suggested that ongoing efforts focused on automating and centralizing data management may aid in further optimization of this methodology for future studies.

Meanwhile, the NIH sponsored STeroids to REduce Systemic inflammation after infant heart Surgery (STRESS) trial [24] has expanded to 18 activated sites, with over 300 patients already enrolled. This ongoing randomized, placebo-controlled, double blind, multi-center trial is designed to evaluate safety and efficacy of perioperative steroids in infants (age < 1 year) undergoing heart surgery with CPB. For this nested “trial within a registry,” all demographic, operative and most of the trial outcomes data are collected using the existing STS-CHSD data collection forms and processes already in place at participating centers. A small subset of additional safety and laboratory variables are not currently collected by the STS-CHSD. These variables will be captured into a separate, limited, web-based study database that will be linked with the STS-CHSD. The STRESS trial overcomes several major hurdles that have previously prevented completion of larger-scale randomized controlled clinical trials in neonates and infants undergoing CPB operations. By leveraging the existing infrastructure of the STS-CHSD, the trial will be conducted with maximal efficiency, and at the lowest possible cost. Both the Pediatric Heart Network’s RLS Study and the National Centers for Advancing Translational Sciences (NCATS) funded STRESS Trial demonstrate the utility of registry data, not just for retrospective observational analyses, but for prospective clinical investigation including clinical trials.

Conclusion

The STS CHSD is a unique resource for research to further our understanding of associations between diagnoses, patient factors, treatment strategies, procedural factors and outcomes in the surgical management and related aspects of the care of patients with congenital heart disease. Summarized herein are reports of STS CHSD-based research published during the past year. Also summarized are ongoing research efforts that use STS CHSD data in combination with data from other sources.

Continuing growth of the STS CHSD and the ongoing development and refinement of tools for quality measurement, together with the numerous pathways for data access by qualified investigators, virtually insures that the STS CHSD will continue to be a resource for research that has the potential, ultimately, to benefit patients by providing evidence and insights that support meaningful progress in patient care.