Abstract
Background
Given improved contemporary survival of adults with congenital heart disease (ACHD), we aimed to evaluate trends in ACHD surgery and outcomes at a single center over a 27-year period.
Methods
Surgical databases were retrospectively queried for patients >18 years old who underwent ACHD surgery between January 1, 1994, and December 31, 2020. A total of 2,195 included patients underwent 2,425 cardiac surgical procedures within the specified time frame. Patients were grouped by era: I, 1994-2000; 2, 2001-2010; and 3, 2011-2020. Trends in primary cardiac diagnosis and surgical management were evaluated.
Results
The median age increased across the eras. The most common primary cardiac diagnoses (n = 2,425) overall were left ventricular outflow tract anomalies (n = 2,019, 83%), atrial septal defect (n = 407, 17%), right ventricular outflow tract anomalies (n = 360, 15%), and ventricular septal defect (n = 110, 4.5%). The most commonly observed procedures overall were operations on the left ventricular outflow tract (n = 1,633, 67%), aorta (n = 675, 28%), coronary arteries (n = 449, 19%), right ventricular outflow tract (n = 323, 13%), and atrial septal defect (n = 264, 11%). Major complications occurred in 10% of cases, and 58 patients died within 30 days of their operation yielding an operative mortality of 2.4%.
Conclusion
To our knowledge, this is the largest single center report on surgery for adults with congenital heart disease. Surgery for ACHD has been performed at our center with relatively low morbidity and mortality over the last few decades.
Keywords
Introduction
Survival rates for patients with congenital heart disease (CHD) continue to improve with adult survival rates reaching >85%. 1 Over the last few decades, the distribution curve representing median age at death of the CHD population has shifted from a bimodal curve that peaked in infancy and adulthood to a unimodal curve that peaks only in adulthood. 2 As of 2000, there are more adults living with repaired or unrepaired CHD (ACHD) in the United States than children. 3 In 2010, ACHD represented 2/3 of the total CHD patient population with 1.4 million patients in the Unites States alone.4,5
Adults with CHD is a broad category comprised of patients with various structural problems ranging from defects discovered and treated primarily in adulthood, to those requiring multiple palliative and/or corrective procedures to reach adulthood with subsequent reinterventions later in life.5,6 The hemodynamic and structural anomalies observed in these patients can lead to various cardiac complications, a need for additional surgical intervention, and noncardiac comorbidities.4,6–8 Thus, patients require lifelong care for CHD,1,3 as well as diseases associated with aging. 7
Our center remains the only tertiary-level center for ACHD in the state of Indiana and, given our long-standing experience with this challenging patient population, we aimed to evaluate postsurgical outcomes and potential trends in surgical management and outcomes over time.
Patients and Methods
This study was approved by Indiana University School of Medicine's Institutional Review Board in 2020 (#2011710163). A preliminary list of 5,901 adult (≥18 years of age) cardiac surgical patients who underwent cardiac surgery between January 1, 1994, and December 31, 2020, with ≥1 potential ACHD diagnosis was generated from electronic and paper medical records. These records were then retrospectively reviewed to identify and enroll 2,195 ACHD patients who underwent cardiac surgery at IU Health University, Methodist, or Riley Children's Hospitals within the specified time frame.
Patients who did not have an ACHD diagnosis were excluded. Patients with an isolated congenitally bicuspid aortic valve (BAV) were included, while patients with a patent foramen ovale were only included if another congenital lesion was present. When not explicitly stated, the presence of a BAV is determined using Figure 1.
Flowchart for determining inclusion of cases involving potential bicuspid aortic valves.
The 2,195 included patients underwent 2,425 cardiac surgical procedures within the study period. Cases were analyzed to identify risk factors for major complications and 30-day mortality. Information about patient demographics, primary cardiac diagnosis, operation, hospital type, congenital training status of the surgeon performing the operation, major complications, and survival was collected. Patient characteristics were used to evaluate the conferred postoperative outcomes of surgical management. Major complications were defined as arrhythmia requiring a pacemaker, unplanned reintervention, ventricular assist device placement, and stroke within 30 days.
The cases were then categorized into three eras based on operative date to allow analysis of trends over the study time frame: Era 1 (1994-2000; n = 180), Era 2 (2001-2010; n = 968), and Era 3 (2011-2020; n = 1,277). Trends in primary cardiac diagnosis and surgical management were also evaluated.
Demographic and clinical characteristics were analyzed to determine if there were differences between the three eras using ANOVAs for continuous variables and χ2 tests for categorical variables. Analyses for the main outcomes of interest were analyzed using χ2 tests. Risk factors for complications and mortality were analyzed using logistic regression methods, starting with bivariate analyses to determine possible risk factors, and then using any that had P < .20 in the bivariate results in a multivariable model to see which remained unattenuated. Survival analyses were performed to generate Kaplan-Meier curves and to determine if there were differences between strata using the log-rank tests. All analytic assumptions were verified. Analyses were performed using SAS v9.4 (SAS Institute).
Results
Table 1 lists patient demographics for the entire cohort and by eras. Median age, proportion of male patients, and weight increased across each era. Overall, 35% (n = 855) procedures were reoperations, and this proportion was similar by era. However, when cases involving isolated BAV were removed from the analysis, the proportion of redo operations significantly increased over time (36%, 50%, and 54% in eras 1, 2, and 3, respectively; P = .007).
Data Presented as Median (Standard Deviation) or Number of Cases (% of Category) Where Appropriate. Patient Demographics.
Primary cardiac diagnoses are listed by frequency in Table 2. The most common diagnoses (n = 2,425) were left ventricular outflow tract (LVOT) anomalies (n = 2,019, 83%), atrial septal defect (ASD) (n = 407, 17%), right ventricular outflow tract (RVOT) anomalies (n = 360, 15%), and ventricular septal defect (n = 110, 4.5%). Diagnoses were not mutually exclusive. Left ventricular outflow tract anomalies primarily included BAV (n = 1,404, 58%) and subaortic stenosis (n = 165, 6.8%). The proportion of BAV increased across all eras (P = .03). Of the 165 cases of subaortic stenosis, 74 were isolated idiopathic hypertrophic subaortic stenosis cases (Table 2).
Primary Cardiac Diagnosis.a
Abbreviations: AVSD, atrioventricular septal defect; D-TGA, dextro-transposition of the great arteries; DORV, double outlet right ventricle; L-TGA, levo-transposition of the great arteries; LVOT, left ventricular outflow tract; PAPVR, partial anomalous pulmonary venous return; PS, pulmonary stenosis; RVOT, right ventricular outflow tract; TGA, transposition of the great arteries; TOF, tetralogy of fallot.
Data presented as number of cases (% of category) of note, coronary artery anomaly designation does not include acquired coronary artery disease.
There was a significant increase in the proportion of left ventricle dominant single ventricle cases observed across the eras (P = .04), though this is likely attributable to a relatively large increase in cases between eras and the small overall sample size involving this lesion (n = 9, 0.4%). Other than the previously identified exceptions, the proportions of primary cardiac diagnoses remained stable across the eras.
The most commonly observed procedures (n = 2,425) were operations on the LVOT (n = 1,633, 67%), aorta (n = 675, 28%), coronary arteries (n = 449, 19%), RVOT (n = 323, 13%), and for ASD (n = 264, 11%) (Table 3). Operative components were not mutually exclusive, as concomitant cases were listed separately in the analysis. Left ventricular outflow tract procedures included aortic valve (n = 908, 37%) and aortic root procedures (n = 727, 30%). The most commonly observed aortic valve procedures were aortic valve replacement (n = 847, 35%) and aortic valve repair (n = 61, 2.5%), while aortic root procedures (n = 727, 30%) primarily consisted of Bentall and Ross procedures (n = 397, 16% and n = 201, 8.3%, respectively). Aortic root enlargement (n = 82, 3.4%) was also a frequently observed LVOT procedure. The proportion of aortic valve procedures increased over time (P < .01) whereas that of aortic root procedures decreased across eras (P < .01), although it is not clear whether these were directly related. Aortic procedures increased across the eras (P < .01) and included primarily replacement/repair of the ascending aorta (n = 611, 25%) and aortic arch (n = 392, 16%). Coronary artery procedures (n = 449, 19%), which peaked in Era 2 (P = .05), included procedures for congenital coronary artery anomalies and for acquired coronary artery disease (Table 3).
Operation Types.a
Abbreviations: AVSD, atrioventricular septal defect; CABG, coronary artery bypass grafting; LV, left ventricle; LVOT, left ventricular outflow tract; MVR, mitral valve replacement; PAPVR, partial anomalous pulmonary venous return; RVOT, right ventricular outflow tract; TVR, tricuspid valve replacement.
Data presented as number of cases (% of category). Table not exhaustive, as operations comprised of <1% of the patient population were excluded.
Fewer than half of the cases in our study were performed by a congenital cardiac surgeon (951/2,425, 39%), and 3.8% of cases were performed jointly by both a congenital and non-congenital cardiac surgeon (91/2,425). This observation is attributable to the inclusion of isolated BAV in our study, as non-congenitally trained surgeons performed a majority of those operations independently (947/1,190, 80%). If those cases were removed from the dataset, congenital cardiac surgeons would have been involved with almost 2/3 of the remaining cases (799/1,235, 65%), performing a majority of them independently (744/1,235, 60%).
Major complications from Era 3 are listed in Table 4 and occurred in 10% of patients. Of the five most commonly observed operations, coronary artery procedures carried the highest rate of major complications (28/203 cases, 14%), followed by LVOT procedures (107/885 cases, 12%). The most frequently observed complications were arrhythmia requiring a pacemaker and return to the operating room for mediastinal bleeding. A higher proportion of LVOT procedures (45/885, 5.1%), coronary artery procedures (9/203, 4.4%), and ASD closures (8/127, 6.3%) resulted in arrhythmias requiring a pacemaker compared with RVOT procedures (2/177, 1.1%) and aortic procedures (10/400, 2.5%). The rates of the identified major complications were otherwise relatively similar across the five most commonly performed procedures. Table 5 lists results of bivariate risk factor analysis. Risk factors for major complications included increasing age, increasing body surface area, operation performed by a non-congenital surgeon, and surgery involving mitral valve replacement/repair, the last of which was significant by multivariable analysis (P = .02).
Outcomes.a
Abbreviations: ASD, atrial septal defect; LVOT, left ventricular outflow tract; RVOT, right ventricular outflow tract; VAD, ventricular assist device.
Analysis of major complications were restricted to Era 3 due to incomplete data in Eras 1 and 2.
Bivariate and Multivariable Logistic Regression Analyses Conducted to Identify Risk Factors for Major Complications.a
Abbreviations: DORV, double outlet right ventricle; LVOT, left ventricular outflow tract; MVR, mitral valve replacement; PAPVR, partial anomalous pulmonary venous return; RVOT, right ventricular outflow tract; TOF, tetralogy of fallot; TVR, tricuspid valve replacement.
Surgery training: “Both” = at least one surgeon with congenital training and at least one surgeon without congenital training performed the case together.
When cases involving isolated BAV were analyzed separately, there was no significant difference in risk of major complications between cases performed by congenital cardiac surgeons and those performed by non-congenital cardiac surgeons by logistic regression (P = .07). The rate of major complications overall was roughly proportional to the number of isolated BAV cases in Era 3, with the exception of stroke within 30 days (Table 6). Isolated BAV cases represented almost 80% of cases involving stroke, with patient ages ranging from 41 to 66 years. Two of the operations were performed for aortic valve endocarditis, 12 cases involved aortic aneurysms, and 1 case involved a surgical ablation of atrial fibrillation. All of the isolated BAV cases resulting in stroke were performed by adult cardiac surgeons.
Outcomes, Isolated BAV Cases.a
Abbreviations: ASD, atrial septal defect; LVOT, left ventricular outflow tract; RVOT, right ventricular outflow tract; VAD, ventricular assist device.
The proportion of major complications that occur in cases with isolated BAV as the primary diagnosis. Analysis of major complications was restricted to Era 3 due to incomplete data in Eras 1 and 2.
Fifty-eight of 2,425 patients died within 30 days of their operation, yielding an operative mortality of 2.4%. As with major complications, coronary artery procedures resulted in the highest operative mortality (29/449 cases, 6.5%), followed by LVOT procedures (47/1,633 cases, 2.9%) (Table 4). Table 7 lists results of bivariate risk factor analysis. Risk factors for operative mortality by bivariate analysis included increasing age, surgery at an adult hospital, and surgery requiring coronary artery procedure, the last of which was significant by multivariable analysis (P < .01). However, cases performed by a congenital surgeon had a reduced risk of operative mortality by bivariate analysis—but not by multivariable analysis. Again, when cases involving isolated BAV were analyzed, there was no significant difference in risk of death between cases performed by congenital cardiac surgeons and those performed by non-congenital cardiac surgeons (P = .40).
Bivariate and Multivariable Logistic Regression Analyses Conducted to Identify Risk Factors for Operative Mortality.a
Abbreviations: DORV, double outlet right ventricle; LVOT, left ventricular outflow tract; MVR, mitral valve replacement; PAPVR, partial anomalous pulmonary venous return; RVOT, right ventricular outflow tract; TOF, tetralogy of fallot; TVR, tricuspid valve replacement.
Surgery training: “Both” = at least one surgeon with congenital training and at least one surgeon without congenital training performed the case together.
Kaplan-Meier analysis revealed that the number of prior procedures did not impact on overall mortality through 25 years of follow-up (P = .70, Figure 2).
Kaplan-Meier analysis evaluating the impact of the number of prior procedures on overall survival, with log-rank test.
Comment
Adult patients with CHD have begun to reach a point where the intersection between congenital heart defects, associated anatomic and physiologic repairs, and age-related conditions can be observed. The age-related disease factors in this population compound the logistically intricate nature of surgically managing ACHD patients. The training of the surgeon, referral for operation at an adult, pediatric, or ACHD specialized center, and the dominant lesion being treated are all factors to consider when caring for these complex patients.
Many ACHD patients require reoperation over the course of their life. 8 Anticipating the likely timing, method, and outcomes of those reinterventions is an increasingly important aspect of caring for this patient population. Prior studies have reported that reoperations comprise 23.1% to 55.2% of the data sets.9,11–14 In this study, 35% of cases were redo operations, with 22% of total cases having ≥1 prior cardiac surgical procedure as a child and 16% with ≥1 prior cardiac surgical procedure as an adult. Interestingly, although the percentage of redo operations in our cohort overall remained relatively stable across the eras (Table 2), when cases involving isolated BAV were omitted, the observed trend in the proportion of redo operations more closely aligned with conventional thought. Our hypothesis is the increasing proportion of isolated BAV cases across the eras combined with the tendency of the lesion to be discovered and surgically remedied in adulthood artificially skewed the analysis of our dataset to favor first time operations. However, bivariate logistic regression analyses show that operative mortality and major complications were not impacted by the number of prior procedures (Tables 5 and 6). Additionally, the number or prior procedures were shown not to have an impact on overall mortality by Kaplan-Meier analysis (Figure 2).
The operative mortality for this study was 2.4%, which is similar to prior reports from single centers (1.5%-2.6%).9,12,13,15 Of the 58 patients who died within 30 days of their operation, 44 underwent an operation by a non-congenital cardiac surgeon (76%), and 57 underwent an operation at an adult hospital (98%). Surgery by a congenital heart surgeon was found to reduce the risk of mortality within 30 days, which was significant by bivariate logistic regression analysis. Kogon et al found that mortality rates were higher when cases were performed by non-congenital cardiac surgeons but identified surgery at a pediatric hospital as a risk factor for mortality. 16 However, surgery at an adult hospital was noted to be a risk factor for 30-day mortality by bivariate logistic regression analysis in our series.
It is our contention that congenital cardiac surgeons as well as their teams at pediatric hospitals are more familiar with the typical recovery pathway of patients with CHD, regardless of age. Thus, they are more prone to recognizing deviation from the anticipated course and proactively adjusting the care paradigm to support a positive result. However, in patients whose defects more closely approximate a more traditional adult cardiac surgery patient (e.g., isolated BAV) or require a concomitant procedure for an acquired lesion such as a coronary artery bypass graft or an aortic procedure, the complication and mortality benefit conferred by the presence of a congenital cardiac surgeon is no longer evident. This finding suggests that if a non-congenitally trained cardiac surgeon were to perform CHD cases more frequently in their practice, the outcomes between congenital and non-congenital cardiac surgeons would tend to equilibrate.
Indiana University has the benefit of a long-standing, accredited ACHD program, aimed at providing excellent outcomes for the growing ACHD population. In the past, the division between cardiac surgical subspecialties was in no way absolute. Thus, congenital cardiac surgeons at our center performing ACHD cases still had a working familiarity with cases which are presently considered to be the domain of non-congenitally trained cardiac surgeons, leading to comparable outcomes between them and their non-congenital counterparts. Moreover, in the current era, our center has benefitted from a close working relationship between the congenital cardiac surgery service and adult cardiac surgery service, permitting an extensive history of combined cases in which a congenital and non-congenital cardiac surgeon were present. This maximizes the strengths of both disciplines, allowing components of a given case which are generally more suited to one service or the other to be completed by a member of that team. As the ACHD population continues to age, it will become increasingly important that either congenital cardiac surgeons who perform ACHD operations maintain the skills required to surgically correct acquired cardiac lesions or that institutions promote combined cases between congenital and non-congenital cardiac surgeons to supply excellent patient outcomes.
One goal of this study was to analyze the trends in the surgical management of ACHD patients. To capture the spectrum of ACHD, cases involving isolated BAV were included in the analyses, regardless of whether another congenital cardiac defect was present. Although inclusion of BAV in the literature has traditionally been contentious, an article published by Nelson et al indicates that the risk of mortality of patients with isolated BAV undergoing ACHD surgery may be better predicted by an adult congenital cardiac surgery risk model than traditional adult cardiac surgery risk models. 8 Thus, it may be time to include isolated BAV in the panoply of congenital cardiac defects studied in adults to allow a more developed understanding of the postsurgical risks faced by the ACHD population as a whole.
Cases involving BAV experienced the only truly significant trend across the eras. From Era 1 to Era 2, the proportion of cases with BAV increased from 49% (n = 89) to 58% (n = 557). This coincides with the inclusion of data from the adult cardiac surgical service, whose database only extended back to Era 2. The increase from Era 2 (n = 557, 58%) to Era 3 (n = 758, 59%) was smaller, supporting the assertion that the increase in frequency of BAV cases observed across the eras is likely the result of including data from the adult cardiac surgical service.
Left ventricular outflow tract anomalies, septal defects, and RVOT anomalies were among the most common diagnoses observed, and the procedures to address them were also among the most frequent operations. These observations were supported by prior studies that reported similar diagnoses and procedures.9–11
This is a single center, retrospective study. While extensive database searches were conducted, it is possible not all relevant cases were captured especially in the earlier eras prior to electronic medical records. Additionally, the database from the adult cardiac surgery service started in 2002, leading to incomplete data capture from 1994 to 2001. Also, complications data were incomplete for Eras 1 and 2, thus analyses were performed using data only from Era 3. Finally, due to the length of the study time frame, very few preoperative data were available consistently across the eras.
To our knowledge, this is the largest single center report of surgery on adults with congenital heart disease, with the next largest identified study being the Mayo Clinic's recent publication on outcomes following reoperative ACHD surgery (N = 1,960). 17 Left ventricular outflow tract anomalies and procedures were the most common across all eras, although the proportions of specific diagnoses and procedures within those categories changed over time. Surprisingly, the number of prior cardiac surgeries did not impact overall survival. Finally, surgery for adults with congenital heart disease has been performed at our center with low rates of major complications and operative mortality over the last three decades.
Footnotes
Abbreviations
Acknowledgments
This project was funded, in part, with support from the Short-Term Training Program in Biomedical Sciences Grant funded, in part by T-35 HL 110854 from the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors would like to thank the Riley Data Outcomes and Research Center, Sloan Goldblatt, and Cameron Colgate for their work on this project.
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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the NIH NHLBI-T35: Short-Term Training Program in Biomedical Sciences, (grant number T-35 HL 110854).
