Research Enterprise of the Congenital Heart Surgeons’ Society 2018 to 2023: History,Accomplishments,Transformation,and Challenges

Abstract

The history of the research arm of the Congenital Heart Surgeons’ Society (CHSS) through 2017 was contained within two prior publications that covered CHSS history in general. The present article is the first to focus explicitly on the research enterprise, with emphasis on the period 2018 to 2023. During this time, the challenges of continuing to build multiple cohorts with lifelong follow-up and to transform the enterprise to a premiere research organization became manifest. Although continuing its commitment to produce impactful research results and to educate the Kirklin/Ashburn Fellow, the research team devoted considerable effort to defining the problems of cohort relevance, workflow, data management, long-term patient follow-up, CHSS member engagement, and the regulatory burden. The team wrote a proposal outlining ways to solve the challenges. A major change from a single-institution “Data Center” to a two-institution Center for Research & Quality (CRQ) was made, assuring increased faculty members and resources. The proposed changes to structure and process began execution in mid-2022. A second Kirklin/Ashburn Fellowship position was created. Between 2018 and 2022, the CRQ produced 17 publications and launched five new research initiatives. This article chronicles the exciting five-year period in which the CHSS research enterprise began a transformation with the intent to become the premiere research organization in the world in the specialty of congenital cardiac surgery.

Keywords

Congenital Heart Surgeons’ Society CHSS history of surgery congenital heart disease

The Congenital Heart Surgeons’ Society (CHSS) Center for Research & Quality (CRQ) is the recently transformed arm of the CHSS charged with carrying out its research mission. Previously called the “Data Center,” the CRQ compiles, follows (ideally, “for life”), and analyzes the largest combined set of cohorts with challenging forms of congenital heart disease in the world. In its 38-year history, it has published seventy articles, presented over 120 abstracts, educated 11 CHSS Kirklin/Ashburn (K/A) Fellows, and introduced novel analytic methods to the specialty of congenital heart surgery. In this article, we review the history of the CRQ, emphasizing the more recent era (2018-2023), and discuss its current and future challenges.

Earlier Era: 1985 to 2017

From casual discussion to research: The development of the CHSS Data Center from its origin through 2017 is chronicled in two prior publications.^1,2 Briefly, the Data Center formed over a decade after a group of 12 surgeons met in 1973 to discuss their complex cases in a collegial setting. In 1984, John W. Kirklin at the University of Alabama (UAB) argued that casual discussion of anecdotal cases alone would not provide sufficient evidence to change and improve clinical practice. Dr Eugene Blackstone (then at UAB) and Dr Kirklin convinced the Society to initiate a formal study of outcomes in patients diagnosed with transposition of the great arteries (TGA). The first article published by this budding UAB-based “Data Center” appeared in the Journal of the American College of Cardiology in 1987. Dr George Trusler was the first author.³ By 1997, five cohorts (the above, plus interrupted aortic arch [IAA], coarctation [COA], aortic valve atresia [AVA], and critical aortic stenosis [AS]) had formed.

Enterprise expansion with a move to Toronto: In 1998, the Data Center moved to the Hospital for Sick Children (“SickKids”) under the leadership of William G. Williams. In 2005, Christopher Caldarone became the first Managing Director. Work Weekends were initiated (Figure 1). Several additional cohorts were created. In 2012, through the initiative of Marshall Jacobs, Julie Brothers, and Chris Caldarone and the generosity of Edward Ludwig and Lisa Nelson Ludwig, the Data Center was funded by the Michael H. Ludwig Memorial Foundation to support research in the anomalous aortic origin of a coronary artery (AAOCA). Research output increased from an average of six articles per five-year period from 1987 to 2006 to ten articles per five-year period from 2006 to 2013. In 2013, William DeCampli succeeded Dr Caldarone as Managing Director. Several changes were made to the structure of the Data Center. First, an Imaging Core Lab (ICL) was created to address the increasing importance of imaging and of the internal consistency of interpretation by concentrating this interpretation among a smaller and more cohesive group of experts who use the same interpretive protocols. Second, the Institutional Data Coordinators (IDC) Group, consisting of coordinators from each CHSS institution, was created to address the challenges of enrollment and follow-up at the institutional level. Third, a formal “Data Center faculty,” each member with specialized expertise, was organized. Luc Mertens from SickKids and Craig Fleishman from Arnold Palmer Hospital were brought on to run the ICL. Linda Lambert, RN, from Primary Children's was recruited to head the IDC Group and lead quarterly IDC conference calls. Anusha Jegatheeswaran, the fifth Kirklin/Ashburn Fellow (now at Great Ormond Street) joined as a clinical as well as data management expert. Fourth, we initiated a 12th cohort, Ebstein’s Anomaly (EBSTEINS) based on a proposal written by three CHSS Members (Drs. Dearani, Knott-Craig, and Pizzaro). At the same time, we discontinued enrollment in the PC cohort to focus on the growth of the other cohorts. We were awarded a grant from the Saving Tiny Hearts Society for research in atrioventricular septal defect, as well as a Children's Heart Foundation grant to support work in the AAOCA cohort. In the five-year period ending in 2017, we published a record number of 15 peer-reviewed articles and presented at over 20 conferences.

Recent Era: 2018 to Present

Data Center transformation to the “Power of Two”: In 2018, we began to look critically at the challenges of (1) “obligatory growth,” that is, the steady increase in the number of “work units” to perform per year (a “work unit” being the processing of an enrollment or a follow-up event) and (2) research productivity. The increase in “work unit” demand was greater than linear with time because it became more laborious to keep patients engaged with the passage of time as they made demographic and health care transitions. Additionally, as the size of cohorts increased and as we requested more data (especially imaging data in the newer cohorts), the burden of data abstraction and dataset production for studies increased. The burgeoning regulation of research (eg, dealing with 83 independent-minded Institutional Review Boards to make cohort-specific amendments) added a third element to the increasing workload. To address the latter problem, we proposed to amalgamate the 11 individual cohorts into a Single Registry Proposal (SRP) that would allow us, in principle, to make structural and process changes across all cohorts by amending a single document. In additional internal meetings, we concluded that the other problems would require substantial restructuring of the Data Center itself and an investment of resources outside its traditional walls at SickKids.

Following discussions at a 2019 Executive Council retreat organized by then-President Joe Dearani, a Data Center Review Committee was created. The committee released a request for proposals in August 2020, authored by Jim Kirklin, Marshall Jacobs, Joe Dearani, and Bill DeCampli, for a “Data Center of the Future.” On December 16, 2021, Joe Dearani and David Overman (then President-Elect) sent a letter to the six authors of the proposal entitled, “The Power of Two: A Joint Proposal by the Cleveland Clinic and the Hospital for Sick Children for a New Congenital Heart Surgery Society Research and Quality Institute” announcing the proposal's selection by the Executive Council.

The “Power of Two” (PO2) proposal was unique in establishing a “Center for Research & Quality” with a “dual” Data Center (renamed the Operations & Coordination Center, “OCC”)—one in Cleveland and the other in Toronto—that could capitalize on resources from two renowned academic healthcare systems and two universities. The total institutional direct financial and “in kind” commitment increased significantly. Additionally, the Data Center faculty now included Drs Hani Najm and Tara Karamlou. The Cleveland Clinic provided funding for a second K/A Fellowship to be based in Cleveland. The second Fellowship itself was a major advance for productivity, as the K/A Fellows have been the workhorses and drivers of most of the CHSS studies. The statistical expertise was strengthened by the tighter engagement with the Clinic's statistical team led by Dr Eugene Blackstone. The OCC faculty now included five members of the statistical editorial staff of the Journal of Thoracic and Cardiovascular Surgery. Michael-Alice Moga, physician director of QA programs in the Labatt Heart Centre at SickKids, was appointed to the faculty as an onsite QA/QI expert. Dr Moga's role is to manage the staff, guide the K/A Fellow, and locally assist with the needs of CHSS QI project leaders in the execution of approved projects. The CRQ leadership structure was determined in conjunction with the Executive Council (Figure 2). The current OCC faculty members are listed in Figure 3.

Equally compelling was the proposal's plan for creating a “hub-and-spoke” structure that would (1) increase CHSS member engagement and (2) further increase research productivity. The “hub” would be the OCC, while the “spokes” would be sets of entities associated with specific expertise or remote investigator teams (Figure 4). Originally conceived by an ad hoc CHSS committee addressing CHSS research, the “hub & spoke” model was to decentralize the conduct of research to make it more practical for independent CHSS investigator teams to conduct research projects from their home institutions, and to capitalize on CHSS institutions’ specialized resources. In meetings of the “joint” OCC faculty in mid-2002, it was decided that activation of the SRP (described earlier in this section) was paramount in creating this structure, as its approval would give the OCC broad regulatory approval to establish data-sharing agreements with CHSS institutions as well as selective outside institutions. This arrangement, in turn, would make datasets available to CHSS investigator teams or members of select “expert” centers outside the OCC to engage in OCC-based studies or even independently conduct studies based on the cohorts. It would also allow for access to a centralized imaging server by a network of “expert” image readers (eg, echocardiographers) that would enable study-specific expert reading to be completed in a fraction of the time previously required. The SRP draft was completed and submitted to the OCC IRBs in March 2023. In the meantime, the OCC began to operationalize the process of producing these datasets and building the ICL network. A memorandum of understanding was approved in December 2022 that allowed the exchange of de-identified data between the OCC campuses.

Additionally, the proposal specified a plan for revamping the recruitment and enrollment of patients, data management, and the follow-up process. The OCC initiated plans to increase patient enrollment at the institutional level by receiving the demographic (“face page”) form used by institutions to log a case into the Society of Thoracic Surgeons (STS) congenital database, then feeding it back to IDCs to encourage enrollment of that patient. This is a “real time” modification of the current model of receiving periodic lists of potentially eligible enrollees based on the CHSS-STS Linkage Project.^4,5

Data flow, data management, and challenges with follow-up, as elucidated in the PO2 proposal, are described in the section “Diagnosis-based cohorts” ahead and in detail in the Online Supplementary Material. A graphical representation of the data flow in the CRQ is shown in Figure 5.

The Data Center had long recognized that cohort enrollment and follow-up would not necessarily need to continue indefinitely, but that reevaluation of the appropriate cohort size and duration of follow-up would warrant periodic assessment. Aside from considering research objectives, this process would help to control the “obligatory growth” of staff workload. For example, the Center decided to cease enrollment into the PC cohort in late 2014. In early 2022, two additional cohorts (AVSD and AAOCA) were evaluated by convening ad hoc “expert” groups of CHSS members. The AVSD group recommended limiting follow-up of AVSD to ten years, reducing the number of abstracted variables, and possibly narrowing the inclusion criteria to “unbalanced” cases. The AAOCA group recommended continuing enrollment and follow-up indefinitely, reducing the variable burden, accelerating image reading, and addressing the biased enrollment and long-term follow-up problem, especially for “medically managed” patients. The OCC plans to evaluate the remaining cohorts in 2023.

In accordance with the Data Center Review Committee's Request for Proposal (RFP) addendum, the PO2 proposal recommended that the Research Committee share some of the “nuts and bolts” responsibilities with the CRQ. The recommendations included (1) recruiting CHSS member engagement and encouraging submission of research or quality proposals, (2) assisting investigators with proposal writing and evaluating final proposals, (3) establishing timelines for proposal execution, (4) requiring and reading investigator progress reports, (5) overseeing investigator plans and progress for meeting presentations and manuscript production, (6) assessing the financial requirement of a proposal and assisting investigators in finding suitable grant opportunities when outside funding is required, and (7) overseeing and providing advice and resources for investigators in the grant writing, submission, and revision process. This plan would have the significant advantage of engaging 5 to 6 more CHSS members in the “hands-on” operational aspects of the research enterprise, thus contributing tangibly to the objective of increased productivity.

A pivotal role of the CHSS in QI/QA was emphasized at the 2019 Executive Council retreat. The CHSS Committee on Quality Improvement and Outcomes sought to integrate with the Research Committee to become the “Quality Leadership” arm of the CRQ. Erle Austin, James Kirklin, and Jim St. Louis spearheaded a QA pilot project tracking one-year outcomes of certain “index” cases, using the UAB/KIRSO organization as its hub. The project is ongoing, with a steadily increasing number of engaged CHSS institutions. The PO2 proposal positioned the OCC as part of a QA “hub-and-spoke” model providing resources for data management and analysis if needed by independent CHSS QA teams working through the Committee on Quality Improvement and Outcomes.

Once the final structure of the CRQ was approved by members of the Data Center Review Committee in June 2022, the OCC began working on putting the provisions of the PO2 proposal into operation. Aside from the tasks listed above, the Center completed the following activities:

Memorandum of Understanding between CHSS, Cleveland Clinic, and SickKids completed: This allowed the Cleveland Clinic and SickKids to begin sharing de-identified data and was the harbinger of the broader Master Participation Agreement.

Single Registry (SRP) Pilot Project: The OCC sent a draft version of the SRP to 12 CHSS institutions for approval by their IRBs. The purpose was to examine the IRB responses so that further changes could be made to the document. The results demonstrated an encouraging breadth of approval of the major points, including proposal adoption without reconsent of already-enrolled patients. After further revisions, the SRP was submitted to the two OCC IRBs in March 2023. Upon approval of the proposal, the SRP was to be distributed to all CHSS institutional IRBs.

Pilot demonstration of technical process of virtual image reading using the Ambra platform (Ambra Health, New York): This was executed between SickKids and Arnold Palmer Hospital. It is the root platform for the ICL's Virtual Image Reader Network.

Selection of the 12^th and 13^th Kirklin/Ashburn Fellows. Justin Robinson, MD, currently in the Integrated 6 program at the University of Maryland matriculated to Cleveland Clinic, and Andres Palacio, MD, a fellow at the Fundacion Cardioinfantil in Bogota began his K/A Fellowship in Toronto, both in July 2023. The history of the Kirklin/Ashburn Fellowship is detailed in the article by Karamlou and coworkers in this issue of the World Journal.⁶

First OCC Work Weekend held at Cleveland Clinic: The OCC faculty convened at the Clinic and attendees participated online. Remarkably, a record number of 83 CHSS members, associates, residents and fellows, IDCs, parent advocacy representatives, and OCC faculty attended some or all of the weekend. The meeting was rife with research ideas as well as constructive suggestions for the projects led by Madison Argo, the 11^th K/A Fellow.

Diagnosis-based Cohorts: The annual growth of cohort enrollment has been cohort- and time-dependent and varied between 5% and 20%. Enrollment status as of September 2022 is summarized in Figure 6. Annual survey-based patient follow-up is active in all cohorts except coarctation. As of January 2023, enrollment is active in 4 of 12 cohorts (shown as an asterisk). EBSTEINS, a newer cohort, is enrolled only at SickKids until activation of the Single Registry proposal. Left ventricular outflow tract obstruction (LVOTO) enrollment was suspended in 2019 and its future status will be determined by an Expert Focus Group conference in 2023. The CRQ is currently considering a proposal to initiate a “corrected” TGA cohort with the earliest start date in 2024.

The ideal number of patients that can be followed as of September 2022 is shown in Figure 7 as the number “presumed alive.” The actual number of patients followed, however, is considerably less than those shown in the figure. (For the active enrolling cohorts, the number not lost to follow-up is shown in the caption.) For the most part, the reasons are either that patients move and cannot be located or that we believe we know their contact, but they either do not respond, or they respond but do not complete and send the data forms. The reasons for the problem are either the loss of accurate contact information as a result of relocation or the lack of a response or an incomplete response from those whose contact information is thought to be accurate. In the PO2 proposal, the OCC proposed several process changes that would possibly ameliorate this problem, including (1) a web-based application (“app”) that pings patients to trigger a response and contains standardized and simplified survey questions, (2) use of commercial apps such as Epic's My Chart (Epic Systems) to allow patients to directly forward annual medical records, (3) incentives, such as periodically sending reports of research progress and patient stories to enrollees, and (4) regulatory changes allowing the OCC to search for “missing” patients using appropriate web-based means. These will be important projects beginning in mid-2023.

Research Accomplishments

Although the appropriate evolution of structure and process enables the CRQ's performance, the metric of success is the research it produces and its impact on our specialty. Despite the myriad of challenges the enterprise faced, research output (measured by number of publications per unit time) steadily increased between 1998 and 2022, as shown in Figure 8. Throughout the most recent five-year era (2018-2022), a Data Center publication appeared on average every 3.3 months.

Although the number of publications may be a metric of enterprise productivity, the cumulative impact of research results is a measure of enterprise value to the specialty. Reed and coworkers have pointed out the challenges of evaluating impact and proposed a methodological framework for evaluating it.⁷ In disciplines such as the sciences a good paper stimulates further research and more papers, making “citation index” a reasonable way to determine impact. In the clinical specialties of medicine, however, a paper's impact is determined by what extent it changes clinical practice. Unfortunately, the latter is infrequently rigorously evaluated. Subjectively, we believe that CHSS research has the potential to be impactful based on several of its previously described features. With this backdrop, what follows is a synopsis and brief commentary of the 18 CHSS publications in the five-year period 2018 to 2022, organized by cohort diagnosis.

Anomalous Aortic Origin of a Coronary Artery

The AAOCA 2019 Working Group studied 285 patients in the AAOCA cohort who either underwent testing for or who otherwise presented with evidence of ischemia⁸ (Figure 9). About 49 (17.2%) of this subgroup had preoperative evidence of ischemia. Those with evidence of ischemia, compared to those with negative ischemia tests, were more likely to have anomalous left coronary, an intramural course, high orifice, or slit-like orifice. Those with anomalous right coronary and ischemia were more likely to have a longer intramural course.

Commentary: The article adds incrementally to the existing evidence and to intuitive notions that an anomalous left coronary, intramural course, high and slit-like orifice are risk factors for clinically significant AAOCA. Ultimately, formal guidelines for AAOCA management (ie, whether to operate or not) should be stratified by anatomic variables, when the risks are better determined using larger sets of advanced imaging data. Additionally, guidelines should ultimately include operative techniques specific to the anatomy, as recently suggested by Jegatheeswaran and DeCampli.⁹

Jegatheeswaran and the AAOCA 2020 Working Group studied 395 patients (58%) who had undergone surgical repair of AAOCA among the 682 patients in the cohort at that time.¹⁰ Mortality was 1%. Overall, 8% of patients developed new moderate or greater aortic insufficiency, 2% developed persistent reduced left ventricular (LV) ejection fraction, and 20% of patients with evidence of preoperative ischemia had evidence of postoperative ischemia.

Commentary: Operation for AAOCA, while conceptually straightforward, is associated with important complications that must be considered in improving operative techniques and overall management strategies for AAOCA. For example, the indications for commissural manipulation should be carefully assessed.

Atrioventricular Septal Defect

Meza, the AVSD Working Group and the CRQ ICL echocardiographers, studied the baseline (preintervention) echo characteristics of 257 patients in the cohort.¹¹ The correlations among three often-quoted indices of “unbalance” (Atrioventricular valve index [AVVI], Right ventricle to left ventricle [RV-LV] inflow angle, and LV inflow index) were weak or moderate. The correlation between each unbalance index and various measures of both atrioventricular valve (AVV) leaflet and ventricular dimensions was also weak to moderate.

Commentary: Suppose one defines “unbalance” as a set of anatomic and physiological characteristics that inform the risk that a complete two-ventricle repair in AVSD will “fail,” that is, result in unacceptable morbidity and mortality. Given that definition, a set of characteristics decisively better than the ones derived originally by Cohen and coworkers could not be determined from the baseline echocardiograms studied.¹² Thus, there remains a wide “gray zone” of AVVI (0.2-0.4) in which “failure” is not well-predicted. Furthermore, the weak correlations among these and numerous other measures reported in the study failed to elucidate the underlying mechanisms by which “failure” occurs. The problem is wide open for further study and the solution continues to be pursued by the CHSS.

Devlin and the Working Group studied the characteristics and outcomes of AVSD patients who underwent pulmonary artery banding (PAB) (50 out of 474 patients [10.5%] in the cohort) as their initial procedure.¹³ Twenty-four infants (5%) had PAB with the intention of staging to a biventricular repair (BVR). Of these, survival at four years after BVR was like that of patients undergoing single-stage complete BVR. The distribution of degree of common AVV regurgitation was similar after PAB compared to the total cohort undergoing BVR.

Commentary: The number of patients in the subgroup of interest was small (24). The reasons surgeons chose to stage to BVR (vs single-stage complete repair) is an important omission and a persistent problem with observational studies using medical record data. The analysis did not support the common belief that PAB will worsen AVV regurgitation.

Callahan and the Working Group studied the association of the presence of an atrial septal fenestration (ASF) after biventricular repair with outcomes in the AVSD cohort.¹⁴ ASF was present in 133 of 581(23%) of total patients. The five-year survival of patients with and without ASF was 83% and 93%, respectively (P = .001). ASF was a significant risk factor for mortality in a multivariable analysis which included adjustment for “unbalance.” Additionally, there was no difference in measures of ventricular unbalance among patients with and without ASF.

Commentary: Surgeon reasons for placing (or leaving) an ASF were unclear from the source data. As mentioned in the prior study, the absence of “reason to treat” data is an important omission and challenging to overcome in nonrandomized observational studies. Consequently, the study provides little guidance for the surgeon as to whether and when an ASF should be employed in AVSD repair. The CRQ is working to solve the data gap.

Critical Left Ventricular Outlet Obstruction

Meza and an LVOT Working Group used unsupervised cluster analysis (Ward's method) to explore classification schemes for critical LVOTO (n = 651 patients) using 136 baseline echocardiographic measures.¹⁵ The cluster algorithm sequentially merged down to three distinct groups. Aortic valve atresia and LVED volume were shown to be significant variables differentiating the three groups. In examining the resulting patient characteristics of the groups, the three groups resembled, but were not identical to the conventional diagnoses of multilevel LV hypoplasia, HLHS, and critical aortic stenosis, respectively.

Commentary: This article was the first to report the use of cluster analysis as a classification scheme in pediatric cardiology and pediatric cardiac surgery using echocardiographic data exclusively. Interestingly, the mitral valve size was not a significant grouping variable, while left ventricular end diastolic (LVED) volume was the dominant differentiator. The results challenge our conventional classification of critical LVOTO purely by mitral and aortic valve echocardiographic metrics. The study allowed us to “get our feet wet” with a machine learning technique that will prove valuable in future CHSS studies.

Devlin and the LVOTO Working Group examined intervention for arch obstruction after the Norwood procedure in 593 patients in the LVOTO cohort;¹⁶ 25% of the patients underwent 218 reinterventions for arch obstruction (168 catheter-based, 50 surgical). Reintervention was associated with tricuspid regurgitation and RV dysfunction. The interdigitation technique of arch reconstruction during the Norwood was protective against reintervention. Surgical arch reintervention (as opposed to catheter-based reintervention) was protective against a subsequent reintervention. Arch reintervention was not associated with mortality.

Commentary: The interdigitation technique was described in 2005 by Burkhart and coworkers in a study based at The Hospital for Sick Children.¹⁷ The study reported a zero incidence of arch obstruction with a median follow-up of three years. David Ashburn, the first Kirklin Fellow, was a coauthor. In this 2019 study, few institutions were consistently using this technique. This study significantly reinforced the recommendation that interdigitation (with complete coarctation resection) be performed in the Norwood operation. It would be interesting for the CHSS to conduct a survey, four years hence, to see whether surgeons have largely adopted the recommended technique. As an aside, an advanced statistical technique, modulated renewal, was used in this study. This technique allows one to “update” risk factors for a prescribed outcome after one or more time-related “interventions” or events. The CRQ's first use of this was in a 2010 study by Jegatheeswaran and coworkers studying reintervention after the repair of IAA.¹⁸

Slieker and a special LVOT Imaging Working Group analyzed “baseline” (preintervention) echocardiographic studies of 651 patients in the LVOT cohort.¹⁹ Remarkably, one “expert reader” performed over 150 morphologic and functional measurements on each echo study, for a total of nearly 100,000 measurements. They compared measurements among several different LVOTO subtypes. The most common conventional subtypes and their frequencies were AA/MA 29%, AA/MS 20%, AS/MS 26%, and isolated AS 18%. Retrograde flow was seen in the ascending aorta in 43% of AS/MS and in 10% of isolated AS. Moderate to severe TR was seen in 13% of cases.

Commentary: The main impact of the article may be to inform congenital cardiac specialties that the CHSS has this robust, high-fidelity dataset. The cases can be linked patient-wise to clinical data from other datasets from the LVOTO cohort, providing a rich source of information for future studies.

Meza and an LVOT Working Group studied the effect of the timing of stage-2 palliation (S2P) on survival from the time of stage-1.²⁰ Using risk factors for survival determined from multivariable analysis, patients were stratified as low, medium, or high risk. Using conditional survival analysis, they determined that S2P performed after age three months was associated with optimal survival in low- and medium-risk patients. High-risk patients had lower survival when S2P was performed before age six months.

Commentary: The study supported the commonly held belief that the performance of S2P is progressively riskier, the younger the age below about three months. The question was important because there was a known constant hazard of complications or death during the stage-1/stage-2 interstage interval, implicitly encouraging earlier S2P. This risk has been largely mitigated with home surveillance programs and catheter-based interventions on the systemic to pulmonary artery shunt, and perhaps the Sano variant itself. Thus, an upper limit on the optimal timing of S2P may be less definite, depending on the variable time that the effects of chronic pulmonary over- or under-circulation become manifest. Some of these authors nearly simultaneously analyzed the Single Ventricle Reconstruction Trial public dataset, asking the same question.²¹ The conclusions were similar.

Argo and an LVOTO working group examined the characteristics of the 214 patients in the LVOTO cohort (accrual 2005-2019) who underwent the hybrid stage 1 (HS1) procedure (bilateral PA bands ± ductal stent);²² 69% had ductal stent and 61% had atrial septostomy. The median follow-up was seven years. Reasons for performing the HS1 varied among institutions. Competing risks analysis showed that, by five years post HS1, 9% had a biventricular repair, 36% a Fontan, 12% a transplant, 35% died, and 8% were alive without transition. Factors associated with death included low birth weight, tricuspid regurgitation prior to HS1, and older age at HS1.

Commentary: The study illustrated broad variations in the initial reasons for performing HS1 as a first intervention, in the subsequent pathways of management, and in pathway-dependent outcomes. It was largely a descriptive study. This work helped to motivate what studies must be done to determine when HS1 is the best starting point for managing patients with critical LVOT. The K/A Fellow (Madison Argo) and LVOTO working groups are currently engaged in follow-up analyses, attempting to get closer to the answer.

Aortic Valve Atresia

Stackhouse and a working group analyzed data from the CHSS AVA cohort (enrollment era 1994-2000)²³; 453 (87%) patients underwent initial surgical palliation and 68 (13%) underwent primary transplantation. At 15 years, the survival was 65% and 40% in the transplantation and surgical palliation groups, respectively. Transplantation survival included 13% waitlist attrition. Functional health status (FHS) was comparable in both groups but lower than that of the general population. The authors concluded that, in this early era, primary transplantation was a better management approach for the high-risk infant with aortic valve atresia.

Commentary: As recognized by the authors as well as by two commentaries that accompanied the article, this study, published in 2020, analyzed data accrued two decades earlier.^24,25 Given remarkable improvements in management and survival in the subsequent two decades, the results of a study of a more contemporary cohort may be different. Nonetheless, it seems still true in 2023 that, while surgical palliation remains a first choice for the average-risk patient, transplantation is considered by most programs as a reasonable, if not the only option for the high-risk patient (typically, one with ventricular dysfunction and/or AV valve regurgitation). Developments in xenotransplantation may further change the equation.

Aortic Valve Stenosis

Devlin and an AVS Working Group evaluated the accuracy of the CHSS Critical Aortic Stenosis Calculator.²⁶ This calculator, featured on the CHSS Data Center website between 2001 and 2015 and based on data accrued between 1994 and 2001, was intended to predict five-year survival difference between biventricular (BVR) and univentricular (UVR) repair for neonates with critical aortic valve stenosis. The basic methodology was based on a CHSS study published by Lofland and coworkers in 2001.²⁷ In the present article, the authors used a subset of 246 critical aortic stenosis cases taken from the LVOTO cohort and enrolled between 2005 and 2013. They predicted the optimal management pathway for this population using the calculator and compared the observed and predicted outcomes. The calculator underestimated survival in both surgical pathways (by about 20%). Of note, 60% of BVR patients were treated discordantly from the “optimal” pathway predicted by the calculator. Finally, discordant management and degree of discordance were not associated with survival.

Commentary: The calculator was deemed no longer predictive, and we removed it from the CHSS website in 2015. The reason was obvious: The data were from a prior era. Both management schemes and outcomes changed significantly two decades later. The CRQ is evaluating the feasibility of “rebuilding” a predictive calculator using the LVOTO cohort. A similar challenge exists for predicting optimal pathways for managing unbalanced complete atrioventricular canal defects. With a large enough cohort, a machine learning approach such as cluster analysis may be a better approach than that based on regression (the methodology used in the AS calculator). Such a model, ideally, could be continuously updated (“taught”) to reflect the most recent data and, therefore, remain relevant.

Interrupted Aortic Arch

Jegatheeswaran and a Working Group evaluated functional health status (FHS) following the repair of IAA.²⁸ The cohort accrual years were 1987 to 1997. The surveyed survivors were ages 13 to 24 years. Child and adult questionnaires were distinct, and a third questionnaire targeting FHS in 22q11 deletion syndrome was used. Domain scores were generally within normal population limits, except for significantly lower scores in the physical functioning domain among adults. Interestingly, in one-third of domains, scores were significantly higher than the population norm. Factors associated with lower scores were 22q11 deletion, reported behavioral and mental health issues, and a greater number of cumulative medical/surgical procedures. Morphologic subtypes and initial repair strategy were not directly associated with lower scores.

Commentary: A 2010 CHSS study of IAA outcomes after initial repair demonstrated a persistent hazard for subsequent interventions and mortality.¹⁸ Overall survival was only 60% at 21 years. The article described IAA as a chronic disease. Aside from mortality, disease impact and the effectiveness of palliative therapy in any chronic disorder should include patient-reported outcome measures (PROMs), including FHS. Thus, the present study is an important contribution to determining the expectation for FHS in IAA patients with intermediate to long-term survival. A further impact is that the article is a reminder that such assessments need to be periodically repeated in more contemporary cohorts. As management strategies and outcomes improve with the era, so should FHS.

Pulmonary Conduit

Callahan and the PC Working Group studied 355 patients who had undergone surgical replacement (“PC2”) of an initial pulmonary conduit to determine the risk factors for a subsequent valve or conduit replacement (PC3).²⁹ The time-related risk of undergoing PC3 was associated with the use of aortic allograft for PC2, a smaller PC2 z-score, a higher z-score for the first (“index,” “PC1”) conduit, younger age at PC2, and concomitant aortic valve intervention at PC2.

Commentary: When possible, surgeons should avoid the use of aortic allografts and should “somewhat” oversize the conduit for PC2. Although this study did not establish an upper limit on the optimal z-score for PC2, prior studies showed that the risk of earlier dysfunction, intervention, or replacement of a conduit increased for a conduit z-score > 2.5 to 3.0 at the time of implantation.

Transposition of the Great Arteries

Devlin and a CHSS Working Group examined long-term survival and FHS in the TGA cohort.³⁰ The PedsQL questionnaire was administered in 2017 to assess FHS. The cohort's accrual period was 1985 to 1990; 830 neonates were enrolled. The median follow-up was 24 years. Survival at 30 years after the repair was 80% for arterial switch, 81% for Mustard, 70% for Senning, and 86% for Rastelli. Patients reported FHS like the general population in all domains except physical health. Arterial switch patients reported higher FHS than atrial switch patients in all domains. In the physical health domain, chest pain, syncope, and permanent pacing were most frequently reported.

Commentary: Multiphase parametric hazard analysis showed a nearly constant hazard for death after two years for both the arterial switch and Senning patients, but an increasing hazard for death for Mustard patients (Figure 10). Practitioners should be wary of this concerning trend in the latter group as they continue to age. It is a mistake to consider the atrial switch “of historical interest only.” Cavaet: All three FHS questionnaires were completed by only 96 patients; Of these, only 22 patients had had an atrial switch. In this small sample, one must be careful in drawing clinical inferences from the mathematical analysis.

Jegatheeswaran and a CHSS Working Group published the results of an FHS questionnaire, the SF-36, administered in 2010 to the TGA cohort.³¹ At that time, 659 of the 868 enrollees (76%) were not known to be dead. Of the 659, 217 patients (33%) responded with 210 valid completed questionnaires. The median age was 23 years. Of these 210 patients, 154 (73%) had previously completed a different FHS survey (the CHQ-CF87) in the year 2000, and 71(34%) completed the PedsQL survey in 2017. (Results of the latter survey were reported by Devlin et al,³⁰ as described above). TGA patients scored higher than population norms in most domains of the SF-36, and similarly in the domains of physical functioning and emotional role. SF-36 domain scores did not depend on cardiac morphology or repair type. As discussed above, 2017 scores in the PedsQl survey were like the population norm, except for the domain of physical health. Likewise, CHQ-CF87 scores from the year 2000 were also like population norms, except for the domain of self-esteem. The investigators concluded that, overall, adult survivors with repaired d-TGA can expect normal FHS.

Commentary: The first CHSS study to examine FHS was that of Culbert and coworkers in 2003 using the TGA cohort.³² All told, the CHSS has conducted FHS surveys six times in four cohorts (TGA, AVA, PA/IVS, and IAA). Five surveys have been interpreted to show that “on the whole” FHS is comparable to the population norm, the possible exception being the domain “physical health” (and in one, “self-esteem”). FHS can be rapidly and inexpensively assessed using short, standardized, and “validated” patient questionnaires that can be mailed to the patient. There are known problems with these surveys, however. These problems are briefly reviewed in this article.³¹ One wonders whether we are getting a realistic picture of the quality of health and life in these patients who have had one or more complex neonatal open-heart operations and face persistent or recurrent problems throughout their life. A more challenging task, but one that would be more impactful, is to combine the questionnaires with exercise stress testing and a battery of neurodevelopmental tests. In the 2013 CHSS study, Karamlou et al actually did gather exercise stress test results with the FHS questionnaires and found a subjective discrepancy between the two modalities of health status assessment.³³

Tricuspid Atresia

Callahan and the TA Working Group examined the outcome of patients with tricuspid atresia/normally related great vessels (TA/NRGV), focusing on survival after a superior cavopulmonary connection (CPC).³⁴ The cohort accrual period for the study was 1999 to 2021. Among 417 patients enrolled in the cohort, 382 (92%) had undergone CPC. By five years following CPC, 5% (17) died or were transplanted without Fontan, 91% transitioned to Fontan and 4% were alive without transition. Associations with mortality included significant preoperative mitral regurgitation (n = 9.2%), mitral valve repair at the time of CPC (six, 1.6%), PA band at the time of CPC (34.9%), number of postoperative CPC interventions (in seven patients, 2%) and CPC takedown (five, 1.3%).

Commentary: The low mortality in the study supports the view that TA/NRGV is apparently the optimal morphological and physiological substrate for a successful transition to Fontan. Mitral regurgitation is an important risk factor, as is significant AV valve regurgitation in any single ventricle patient. Fortunately, its incidence in TA/NRGV is very low. 9% of patients underwent PA band at the time of CPC. It was associated with mortality. One can only speculate why these patients needed an accessory source of blood flow. In the contemporary era, clinicians have become more adept at managing pulmonary blood flow in the interstage and, barring unremediable pathology in the lungs or pulmonary vasculature, can “prep” the pulmonary vasculature for a successful CPC alone without an accessory source of blood flow across all single ventricle diagnoses.

Perspectives and Challenges

In August 2021, the National Heart, Lung, and Blood Institute conducted a workshop on the future of pediatric cardiovascular research.³⁵ Overall, the key directions recommended by the group were (1) studying the impact of CHD across the life span, (2) developing new outcome measures relevant to the patient and with robust event rates, (3) performing more studies whose results can improve clinical practice, and providing pathways to implement the results into practice, (4) supporting “discovery” science (basic, transformational, or what Thomas Kuhn called “paradigm science”), (5) building collaborative scientific networks from existing individual networks and registries, and (4) enriching diversity.

Where does CHSS research lie in the context of these directions? First, it should be appreciated that several of the objectives recommended by the NHLBI workshop were already put into operation and utilized in CHSS research, most importantly the “lifespan disease framework” (long-term periodic follow-up) and inclusion of “alternative” outcome metrics that can be assessed robustly and repeatedly such as structural and functional imaging measurements, interventions, and PROMs (at least FHS). Additionally, our study results (some summarized in this article) have represented a balance between discovery and observational trials capable of clinical implementation. As a general guideline, we should continue to focus on studies that address health issues across the life span that may be influenced by surgical management, whose designs are minimally flawed, whose results are reliable, and whose implications can be clinically implemented. Although long-term follow-up is the root of the life span framework, long-term enrollment enables us to assess implementation of prior study recommendations and to determine evolution/progress in clinical management. Although the CHSS bylaws do not explicitly limit CHSS research to “surgical” research, it would make sense for us to continue to emphasize “surgically relevant” problems such as operations, techniques, imaging, and modeling guidance (including computational dynamics and 3D print models), complications, protocols, guidelines, and comparative outcomes of surgical strategies. For example, we should look more at prediction algorithms that inform preferable surgical strategies and are easily implementable into clinical practice.

The PO2 proposal focused on what tools and methodologies we have not developed or applied and must do so to fulfill our own research objectives. First, it is paramount for us to gain CHSS institutional approval of our Single Registry proposal this year to reduce the regulatory burden that constrains our research capability and efficiency. Second, we must solve the follow-up problem by leveraging web-based communication apps, patient/parent advocacy groups, and the ability of patients to directly control sharing of their medical records. Without these solutions, none of the other objectives can be achieved because, in short, we will not be able to acquire the requisite source data to do so. It is important to note that the bold transformation outlined in the PO2 proposal has been activated for barely 18 months from the time of this publication. It remains to be demonstrated that the structures and processes described will actually work to expand and improve the CHSS research enterprise. Assuming they work, however, the CHSS will be poised to address other yet-uncharted elements recommended by the NHLBI workshop, such as:

Should the CRQ enable and promote prospective studies? Surgically relevant adaptive or pragmatic trials?

Should the CRQ engage in selected surgically relevant device or graft trials?

Should the CRQ join, partner, or collaborate with other pediatric cardiac research networks such as the Congenital Cardiac Research Collaborative or Cardiac Networks United? Or in more focused registries such as the Fontan Outcomes Network (FON) registry?

Should the CRQ establish an open data-sharing agreement with the STS Congenital Database and partner in studies?

How can the CRQ further improve data management to enable growth toward “big data” and application of machine learning techniques?

It is by no means necessary to implement all the NHLBI elements into the CHSS research mission, no more than it is to maximize the number of articles published. The CRQ must pick its battles, that is, its areas of interest and expertise and focus on research quality, impact, and implementation. As the premiere organization of our specialty, the CHSS has the resources, skills, ingenuity, and drive to accomplish these lofty goals.

Supplemental Material

sj-docx-1-pch-10.1177_21501351231176219 - Supplemental material for Research Enterprise of the Congenital Heart Surgeons’ Society 2018 to 2023: History, Accomplishments, Transformation, and Challenges

Supplemental material, sj-docx-1-pch-10.1177_21501351231176219 for Research Enterprise of the Congenital Heart Surgeons’ Society 2018 to 2023: History, Accomplishments, Transformation, and Challenges by William M. DeCampli in World Journal for Pediatric and Congenital Heart Surgery

Footnotes

Acknowledgments

The author thanks all the CHSS leaders and members who are passionate about quality, research, and innovation in our specialty and who consistently provided advice over the last ten years. The directives described in this article are, in large part, a result of these interactions. On behalf of the CHSS, the author also thanks donors and supporters who have contributed to CHSS research efforts. Finally, the author wishes to acknowledge William G. Williams, founding member of the CHSS, and for 25 years Executive Director of the Data Center. Dr Williams—unassuming but always present; knowledgeable but practical; leading by giving others the lead—has been the beacon guiding the Data Center through changing times.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

William M. DeCampli

Supplemental Material

Supplemental material for this article is available online.

Abbreviations

References

Mavroudis

Williams

. History of the Congenital Heart Surgeons Society. World J Pediatr Congenit Heart Surg. 2015;6(4):541‐550.

Mavroudis

Kirklin

DeCampli

. Incremental history of the Congenital Heart Surgeons’ Society (2014-2018). World J Pediatr Congenit Heart Surg. 2018;9(6):668‐676.

Trusler GA, Castaneda AR, Rosenthal A, Blackstone EH, Kirklin JW. Current results of management in transposition of the great arteries, with special emphasis on patients with associated ventricular septal defect. J Am Coll Cardiol. 1987;10(5):1061-1071.

Jacobs

Pasquali

Austin

, et al. Linking the congenital heart surgery databases of the Society of Thoracic Surgeons (STS) and the Congenital Heart Surgeons’ Society (CHSS): part 1—rationale and methodology. World J Pediatr Congenit Surg. 2014;5(2):256‐271.

Jacobs

Pasquali

Austin

, et al. Linking the congenital heart surgery databases of the Society of Thoracic Surgeons (STS) and the Congenital Heart Surgeons’ Society (CHSS): part 2—lessons learned and implications. World J Pediatr Congenit Surg. 2014;5(2):272‐282.

Karamlou T. History of Kirklin-Ashburn fellowship. World J Pediat Congenital Heart Surg. 2023.

Reed

Ferre

Martin-Ortega

, et al. Evaluating impact from research: a methodological framework. Res Policy. 2021;50(4):104147. (1-14). https://doi.org/10.1016/j.respol.2020.104147.

Jegatheeswaran

Devlin

McCrindle

, et al. Features associated with myocardial ischemia in anomalous aortic origin of a coronary artery: a Congenital Heart Surgeons’ Society study. J Thorac Cardiovasc Surg. 2019;158(3):822‐834.

Jegatheeswaran

DeCampli

. Toward more granular guidelines in AAOCA: associating anatomical details with specific surgical strategies. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2023;26:63‐74.

10.

Jegatheeswaran

Devlin

Williams

, et al. Outcomes after anomalous aortic origin of a coronary artery repair: A Congenital Heart Surgeons’ Society Study. J Thorac Cardiovasc Surg. 2020;160(3):757‐771.

11.

Meza

Devlin

Overman

, et al. The Congenital Heart Surgeons’ Society complete atrioventricular septal defect cohort: baseline, preintervention echocardiographic characteristics. Semin Thorac Cardiovasc Surg. 2019;31(1):80‐86.

12.

Cohen MS, Jegatheeswaran A, Baffa JM, et al. Echocardiographic features defining right dominant unbalanced atrioventricular septal defect: a multi-institutional Congenital Heart Surgeons' Society study. Circ Cardiovasc Imaging. 2013;6(4):508-513.

13.

Devlin

Jegatheeswaran

McCrindle

, et al. Pulmonary artery banding in complete atrioventricular septal defect. J Thorac Cardiovasc Surg. 2020;159(4):1493‐1503.

14.

Callahan

Jegatheeswaran

Barron

, et al. Congenital Heart Surgeons’ Society Atrioventricular Septal Defect Working Group. Association of atrial septal fenestration with outcomes after atrioventricular septal defect repair. J Thorac Cardiovasc Surg. 2022;163(3):1142‐1152.

15.

Meza

Slieker

Blackstone

, et al. Congenital Heart Surgeons’ Society. A novel, data-driven conceptualization for critical left heart obstruction. Comput Methods Programs Biomed. 2018;165:107‐116.

16.

Devlin

McCrindle

Kirklin

, et al. Intervention for arch obstruction after the Norwood procedure: prevalence, associated factors, and practice variability. J Thorac Cardiovasc Surg. 2019;157(2):684‐695.

17.

Burkhart

Ashburn

Konstantinov

, et al. Interdigitating arch reconstruction eliminates recurrent coarctation after the Norwood procedure. J Thorac Cardiovasc Surg. 2005;130(1):61‐65.

18.

Jegatheeswaran A, McCrindle BW, Blackstone EH, et al. Persistent risk of subsequent procedures and mortality in patients after interrupted aortic arch repair: A Congenital Heart Surgeons' Society study. J Thorac Cardiovasc Surg. 2010;140(5):1059-1075.

19.

Slieker

Meza

Devlin

, et al. Pre-intervention morphologic and functional echocardiographic characteristics of neonates with critical left heart obstruction: A Congenital Heart Surgeons Society (CHSS) inception cohort study. Eur Heart J Cardiovasc Imaging. 2019;20(6):658‐667.

20.

Meza

Hickey

McCrindle

, et al. Congenital Heart Surgeons’ Society. The optimal timing of stage-2 palliation after the Norwood operation. Ann Thorac Surg. 2018;105(1):193‐199.

21.

Meza

Hickey

Blackstone

, et al. The optimal timing of stage 2 palliation for hypoplastic left heart syndrome: an analysis of the pediatric heart network single ventricle reconstruction trial public data set. Circulation. 2017;136(18):1737‐1748.

22.

Argo

Barron

Eghtesady

, et al. Initial bilateral pulmonary artery banding for infants with critical left heart obstruction: a multi-institutional analysis of outcomes. Circulation. 2022;146(Suppl 1). https://www.ahajournals.org/doi/abs/10.1161/circ.146.suppl_1.10158

23.

Stackhouse

McCrindle

Blackstone

, et al., The Congenital Heart Surgeons Society. Surgical palliation or primary transplantation for aortic valve atresia. J Thorac Cardiovasc Surg. 2020;159(4):1451‐1461.

24.

Mascio

. Commentary: unicorns and leprechauns. J Thorac Cardiovasc Surg. 2020;159(4):1462‐1463.

25.

Mitchell

. Commentary: no justification for transplantation in place of standard surgical palliation. J Thorac Cardiovasc Surg. 2020;159(4):1463‐1465.

26.

Devlin

Hickey

Morgan

, et al. Evaluating the Congenital Heart Surgeons’ Society (CHSS) critical aortic stenosis calculator in a new patient cohort. Originally published 5 Nov 2018. Circulation. 2018;138:A14267.

27.

Lofland

McCrindle

Williams

, et al. Congenital Heart Surgeons Society. Critical aortic stenosis in the neonate: a multi-institutional study of management, outcomes, and risk factors. J Thorac Cardiovasc Surg. 2001;121(1):10‐27.

28.

Jegatheeswaran

Jacobs

Caldarone

, et al. Self-reported functional health status following interrupted aortic arch repair: A Congenital Heart Surgeons Society study. J Thorac Cardiovasc Surg. 2019;157(4):1577‐1587.

29.

Callahan

Jegatheeswaran

Blackstone

, et al. Time-related risk of pulmonary conduit re-replacement: A Congenital Heart Surgeons Society study. Ann Thorac Surg. 2022;113(2):623‐629.

30.

Devlin

Jegatheeswaran

Williams

, et al. Late survival and patient-perceived health status of the Congenital Heart Surgeons’ Society dextro-transposition of the great arteries cohort. Ann Thorac Surg. 2019;108(5):1447‐1455.

31.

Jegatheeswaran

Devlin

DeCampli

, et al. Longitudinal functional health status in young adults with repaired dextro-transposition of the great arteries: a congenital heart surgeons’ society study. J Thorac Cardiovasc Surg. 2020;159(2):604‐614.

32.

Culbert

Ashburn

Cullen-Dean

, et al. Congenital Heart Surgeons’ Society. Quality of life of children after repair of transposition of the great arteries. Circulation. 2003;108(7):857‐862.

33.

Karamlou

Poynter

Walters

, et al. Long-term functional health status and exercise test variables for patients with pulmonary atresia with intact ventricular septum: A Congenital Heart Surgeons Society study. J Thorac Cardiovasc Surg. 2013;145(4):1018‐1027.

34.

Callahan

Jegatheeswaran

Barron

, et al. Congenital Heart Surgeons’ Society Tricuspid Atresia Working Group. Factors associated with mortality or transplantation versus Fontan completion after cavopulmonary shunt for patients with tricuspid atresia. J Thorac Cardiovasc Surg. 2022;163(2):399‐409.

35.

Opotowsky

Allen

Bucholz

, et al. Pediatric and congenital cardiovascular disease research challenges and opportunities. J Am Coll Cardiol. 2022;80(23):2239‐2250.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB