Ann Thorac Surg 2007;83:1438-1445
© 2007 The Society of Thoracic Surgeons
Original Articles: Cardiovascular
Causes of Death After Congenital Heart Surgery
Marsha Ma, MDa,
Kimberlee Gauvreau, ScDb,
Catherine K. Allan, MDb,
John E. Mayer, Jr, MDc,
Kathy J. Jenkins, MD, MPHb,*
a Tufts University School of Medicine, Children’s Hospital Boston, Harvard Medical School, Boston, Massachusetts
b Department of Cardiology, Children’s Hospital Boston, Harvard Medical School, Boston, Massachusetts
c Department of Cardiovascular Surgery, Children’s Hospital Boston, Harvard Medical School, Boston, Massachusetts
Accepted for publication October 27, 2006.
* Address correspondence to Dr Jenkins, Department of Cardiology, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115 (Email: kathy.jenkins{at}childrens.harvard.edu).
 |
Abstract
|
|---|
Background: There has been little research about the causes of death after congenital heart surgery.
Methods: To determine whether mode of death differs after congenital heart surgery, we evaluated the cause of death for 100 consecutive postoperative deaths at our institution. Mode of death was determined based on retrospective chart review including available autopsy reports. Low output states were categorized into ventricular failure; inadequate postoperative physiology (technically adequate surgery and ventricular function, but persistent low cardiac output); pulmonary hypertension; and atrioventricular valve regurgitation.
Results: There was considerable anatomic diversity among patients who died; 46 patients had single-ventricle physiology. The vast majority of patients (n = 79) were in the intensive care unit before surgery. Surgical repairs were revised at initial operation in 22 cases; 7 patients died in the operating room. Seventy-three patients had technically adequate surgical procedures, 23 had residual anatomic defects, and 4 were indeterminate. Thirty patients underwent additional surgical and 9 catheter-based procedures, although some were classified as rescue procedures performed to address minor anatomic or physiologic abnormalities as a last hope to rescue the patient from impending demise. Of 100 deaths, most (n = 52) were due to low cardiac output: 24 inadequate postoperative physiology, 19 ventricular failure, 8 pulmonary hypertension, and 1 valvar regurgitation. Other significant causes of death included sudden cardiac arrest (n = 11), sepsis (n = 11), and procedural complications (n = 8).
Conclusions: More than half of the deaths were due to low cardiac output, but not exclusively ventricular failure.
|
Introduction
|
|---|
Despite recent improvements in survival [1], congenital heart disease remains an important cause of death in childhood. Since correction of serious congenital heart defects generally requires a major surgical procedure, the perioperative period remains a vulnerable time for patients. Patients who survive with repaired anatomy can often achieve nearly normal health status [2]. Thus, efforts to reduce surgical mortality remains important to reduce the overall burden of these malformations.
Despite the importance of these issues, formal study of how to achieve optimum outcomes is difficult. In addition to the technical challenges of dealing with small structures, complex anatomy, and considerable anatomic diversity, intraoperative management is complex, requiring cardiopulmonary bypass, hypothermia, and often brief periods of circulatory arrest. Study of specific aspects of the procedure, such as depth of hypothermia [3] or target dilutional hematocrit [4], require randomized clinical trials that can take years to complete, and can be hampered by lack of consensus or standardization of other aspects of the procedure.
We sought to identify specific areas where improvement efforts would likely have the greatest impact on mortality reduction after congenital cardiac surgical procedures. To accomplish this, we performed a modified "failure mode effects analysis" for a group of congenital heart operations that did not achieve a favorable outcome and were associated with death during the operative or perioperative period.
|
Material and Methods
|
|---|
The study was approved by the Children’s Hospital Boston Institutional Review Board and informed consent was waived. The most recent 100 consecutive in-hospital deaths after congenital heart surgery at Children’s Hospital Boston before December 31, 2002, extending back to 1995 were identified by review of departmental electronic databases. In-hospital death was defined as any death occurring after the onset of the cardiac surgical procedure, but before hospital discharge. Patients were excluded if they died after discharge. Congenital heart surgery was defined as any surgical procedure for a cardiac defect present at birth, including planned reoperations and revisions of prior procedures. Patients undergoing heart transplant or noncongenital heart surgery, and premature patients (< 37 weeks) whose only surgical procedure was ligation of a patent ductus arteriosus were excluded.
We retrospectively reviewed all preoperative, intraoperative, and postoperative data including autopsy results, when available. Preoperative data collected included cardiac anatomic diagnosis, comorbidities, medications, history of arrhythmia, previous cardiac surgery, cardiac catheterization done within 2 months before the surgery, and surgical priority (elective/inpatient–cardiac ward/inpatient–intensive care unit [ICU]). Preoperative cardiac function was classified based on echocardiographic or magnetic resonance imaging reports.
Intraoperative information recorded included the type of surgical procedure and any intraoperative revisions. An intraoperative revision was defined as a surgical revision that was performed after an initial unsuccessful attempt to separate the patient from cardiopulmonary bypass. Indications for intraoperative revisions included a hemodynamically significant residual defect or a decision to qualitatively change the type of procedure. Typically, these qualitative changes involved conversion from a two-ventricle approach to a single-ventricle approach. Other intraoperative information collected included total bypass time, cross-clamp time, and circulatory arrest time.
Postoperative data recorded included use of cardiac pacing, cardiopulmonary resuscitation, extracorporeal membrane oxygenation, and additional diagnostic or interventional catheterizations or surgical interventions. Information about postoperative anatomy and function based on echocardiography were collected both immediately postoperatively and before death.
The technical success of the surgical procedure to achieve the intended goal of the procedure was classified by a board-certified pediatric cardiac surgeon (J.E.M) into one of the following categories: (1) residual anatomic defect (defined as a correctly planned surgery but technically inadequate because of a residual anatomic problem); (2) adequate surgical execution; or (3) indeterminate. If the procedure was categorized into residual anatomic defect, notation was made as to whether the surgical procedure was then revised. Within the category of adequate surgical execution, the subgroups included incorrect preoperative diagnosis, poor preoperative condition, failed therapeutic plan, and correct therapeutic plan. Failed therapeutic plan was defined as an adequately performed surgical procedure, but with the benefit of retrospective analysis, an alternative nontransplant therapeutic plan may have been preferable. Typically, these cases involved decisions regarding "single-ventricle" versus "two-ventricle" surgical approaches, but this category also included cases in which the choice of initial surgical operation or decision regarding the timing of intervention appeared incorrect, and patients with difficult anatomy. Difficult anatomy was defined as cardiac anatomy that despite adequate surgical execution and uncomplicated postoperative care, the only treatment, in retrospect, was early cardiac transplantation. It was also noted whether the physicians recognized the necessary change and revised the plan.
During the data collection process, we recognized that many patients underwent attempts to correct or revise minor anatomic problems, in the hope that improvements might reverse a clinical course of deterioration. To distinguish these types of "rescue" procedures from those that were done to correct a significant residual anatomic defect from an inadequate technical repair or a misdiagnosis, a pediatric cardiologist (C.K.A.) classified all catheter-based interventions and a pediatric cardiac surgeon (J.E.M.) classified all surgical interventions as major revisions or rescue procedures. Major revisions were defined as corrections of technically inadequate surgical procedures, whereas rescue procedures were defined as procedures undertaken to improve minor residual defects that could incrementally improve cardiac physiologic parameters if corrected.
Mode of death was determined by a board-certified pediatric cardiologist (K.J.J.) and medical student (T.M.M.) by examining the chronologic order of events beginning immediately postoperatively leading to death using all relevant information available to the treating physician including ICU data, echocardiograms, catheterizations, and operative reports. In most cases, the treating physicians had written summaries of their clinical interpretation of the events that occurred at the time of death in the medical record. Deaths were classified according to the clinical impression of the treating physicians, except in cases where unexpected new information was available through autopsy.
The primary mode of death was defined as the initial clinical event that ultimately led to a sequence of events culminating in death, which was not necessarily the same as the preterminal event occurring immediately before death. Any other causes that may have contributed to the death were classified as secondary or tertiary causes. Similar modes of death were grouped together into categories. Etiology of low cardiac output was further classified into subcategories (1) ventricular failure, (2) inadequate postoperative physiology, (3) pulmonary hypertension, and (4) atrioventricular valve regurgitation documented by echocardiogram as severe. Ventricular failure was defined as moderate or severe ventricular dysfunction based on echocardiography; the designation of "inadequate postoperative physiology" was chosen for patients who had a technically adequate repair and no more than mild systemic ventricular dysfunction, but had a persistent low cardiac output state. This "inadequate postoperative physiology" category includes patients with right ventricular dysfunction associated with small left sided structures and secondary high pulmonary pressure as well as those with palliated single ventricle in whom adequate systemic cardiac output could not be sustained. Pulmonary hypertension was identified when elevated pulmonary artery pressures were documented or when there was a clinical picture consistent with pulmonary hypertension and a specific therapeutic intervention to reduce pulmonary vascular resistance. Additional categories for mode of death included arrhythmia, sudden bradycardiac arrest defined as unexpected bradycardia and hypotension requiring cardiopulmonary resuscitation, stroke, hemorrhage, respiratory infection, respiratory insufficiency, sepsis, death due to complications of surgical or catheter procedure, and unknown cause of death.
|
Results
|
|---|
Patient characteristics are summarized in Table 1. Among the 100 patient deaths analyzed, there was considerable variation in the cardiac diagnoses, although many patients had various forms of single-ventricle physiology (n = 46), most frequently hypoplastic left heart syndrome (n = 26). Most deaths occurred in infants less than 1 year of age (n = 81), usually in the newborn period of 30 days or less (n = 60). Prematurity was the most prevalent comorbidity, occurring as a sole comorbidity in 20 of 43 patients and as one of two or more comorbidities in 7 patients. The majority of patients were in the cardiac ICU before their surgical procedure (n = 79); only 9 deaths occurred after elective procedures. The median age at death was 47 days (range, 2 days to 47 years). The median time from surgery to death was 10 days (range, 0 to 155).
Intraoperative information is summarized in Table 2. Seven patients died in the operating room, 3 after an attempted revision. Interestingly, 2 of these patients had pulmonary atresia with right ventricular dependent coronary circulation. Intraoperative revisions also occurred frequently (n = 22) among the 93 patients who initially survived to the ICU.
The method of categorization for the technical outcome of the procedures is summarized in Figure 1
as a flow diagram. Of the 100 patients in the study, 73 were judged to have had a technically adequate surgical procedure, 23 had an important residual anatomic abnormality, and the status of the repair could not be determined in 4. The 73 patients with an adequate surgical procedure were further categorized into those with a correct initial surgical plan (n = 34) and those with a failed therapeutic plan (n = 39). The patients with a failed therapeutic plan were further divided into difficult anatomy (n = 12), patients whose failed therapeutic plan was recognized and revised (n = 14), other (n = 2), those with poor preoperative status due to a complication of a previous interventional procedure (n = 8), and preoperative incorrect diagnosis (n = 3). Of the 23 patients with residual anatomic defects, 19 had revisions in an attempt to address the residual defect, and 4 were not revised.
View larger version (18K):
[in this window]
[in a new window]
|
Fig 1. Flow diagram displays the categorization of the 100 patients in this study into the various categories of technical adequacy. (Preop = preoperative.)
|
|
Postoperative data for the 93 patients who survived to the ICU are summarized in Table 3. Sixty-four patients (69%) did not undergo sternal closure at the time of the initial operation. Twenty (22%) were placed on extracorporeal membrane oxygenation in the operating room. Most of the patients remained critically ill throughout their postoperative course. Only 25 (27%) were ever extubated. Sixty-two patients (67%) required cardiopulmonary resuscitation, and a total of 42 (45%) were treated with extracorporeal membrane oxygenation. Preoperative, postoperative, and predemise evaluation of cardiac function are shown in Figure 2. Statistical significance of change over time was determined using the Wilcoxon signed-rank test. Right ventricular function was significantly worse between the preoperative and predemise time points (p = 0.03) as was left ventricular function (p = 0.02).
View larger version (28K):
[in this window]
[in a new window]
|
Fig 2. (A) Progression of right ventricular function preoperatively, postoperatively, and predemise. Ventricular function was determined by echocardiography. Right ventricular function was statistically significantly worse between the preoperative and predemise time points (p = 0.03). (B) Progression of left ventricular function preoperatively, postoperatively, and predemise. Ventricular function was determined by echocardiography. Left ventricular function was statistically significantly worse between the preoperative and predemise time points (p = 0.02).
|
|
Thirty patients underwent additional cardiac surgical procedures, 8 of which were classified as major procedures and 22 as rescue procedures (Table 4). Postoperative catheterizations were performed in 30 patients; 21 were diagnostic and 9 were interventional. Among the 9 patients undergoing interventional, 12 major interventions were performed on 8 patients, and 2 rescue procedures were performed on 1 patient. Of the 12 major catheter-based interventions, 6 were judged to be adequate to relieve the anatomic problem, and 6 were unsuccessful. Overall, 14 of 93 patients underwent a major reintervention (6 surgical, 6 catherizations, 2 both), and 23 others underwent a rescue intervention.
The mode of death among the entire cohort of 100 patients is shown in Figure 3. Low cardiac output state was further categorized as inadequate postoperative physiology (n = 24), ventricular failure (n = 19), pulmonary hypertension (n = 8), or severe atrioventricular valve regurgitation (n = 1). Other causes of death included sudden bradycardiac arrest (n = 11), sepsis (n = 11), procedural complication resulting in death (n = 8; see Table 5), and respiratory insufficiency (n = 3).
View larger version (34K):
[in this window]
[in a new window]
|
Fig 3. Causes of death after congenital heart surgery in our study population. Low output state includes venticular failure, inadequate postoperative physiology, pulmonary hypertension, and severe atrioventricular valvar regurgitation (AVVR). The category "other" includes causes of death for 1 patient each, n = 100.
|
|
A subcategory of the patients who were ever extubated (n = 25) were further analyzed for their cause of death. Five patients died of inadequate postoperative physiology, 4 of sepsis, 3 of sudden bradycardic arrest, 3 of pump failure, 3 of neurologic injury, and 1 of each of the following: pulmonary hypertension, hemorrhage, necrotizing enterocolitis, respiratory infection, procedural complication, acute aortic dissection, and severe atrioventricular valve regurgitation.
The causes of death for correct surgical plan, failed therapeutic plan, and residual anatomic defects are shown in Figure 4A, B, and C, respectively. The most frequent cause of death was cardiac arrest among patients with a correct surgical plan (n = 8 of 34), inadequate postoperative physiology for patients with failed therapeutic plan (n = 14 of 39), and ventricular failure for patients with residual anatomic defects (n = 7 of 23).
View larger version (58K):
[in this window]
[in a new window]
|
Fig 4. (A) Cause of death for patients with a correct therapeutic plan, n = 34. (B) Cause of death for patients with a failed therapeutic plan, n = 39. (C) Cause of death for patients with residual anatomic defects, n = 23.
|
|
|
Comment
|
|---|
This study was undertaken in an attempt to identify areas of priority to reduce mortality after congenital heart operations, and to examine whether modes of death after congenital heart surgery differ from those occurring after cardiac operations in adult populations. We found that most patients who died were infants with congenital heart defects. Also, most patients either died in the operating room or remained critically ill for their entire postoperative course and never achieved cardiovascular stability. Similar to adult cardiac patients [5] and to the one prior study in congenital cardiac surgery patients [6], a substantial number of infants and children who died from a low cardiac output state had ventricular failure. However, unlike adult patients, a larger number of infants and children who had a technically adequate repair died of low cardiac output state, but maintained good ventricular function. Death in these patients, defined as inadequate postoperative physiology, occurred because the overall cardiovascular physiology, even with a technically adequate repair and intensive support, was inadequate to meet physiologic demands. With the benefit of retrospective analysis, a successful outcome for these cases could only have been achieved by employing a different strategy of care, such as transplantation or a major difference in operative approach. For example, since many of the deaths were of infants with hypoplastic left heart syndrome undergoing classic Norwood operations, it is conceivable, although unproved, that altered surgical techniques such as the use of right ventricle to pulmonary artery conduit instead of the Blalock-Taussig shunt could have achieved a more stable physiology.
Interestingly, there were a few patients with pulmonary atresia, intact ventricular septum, and complete aortocoronary atresia with all coronary blood supply originating from the right ventricle, which we now understand is associated with essentially 100% mortality and for which cardiac transplant is probably the only treatment possibility [7]. The limited availability of neonatal cardiac donors may have influenced the decision to proceed with surgical palliation rather than immediately listing for transplantation despite the high risk of mortality. These infants are an indication of the need for continued careful reassessment and choice of therapeutic strategy.
Importantly, most infants and children who died were judged to have had a technically satisfactory repair without major procedural complication. These findings contradict the belief that technical surgical problems are currently the major factor responsible for mortality after congenital heart operations. In this study, however, we found that despite an adequate surgical procedure, there were still a number of patients (16) with borderline anatomy, defined as anatomy that was not clearly requiring a single-ventricle or two-ventricle repair, that in retrospect should have had a different initial plan. This recognition occurred in the operating room after direct visualization of anatomy or postoperatively when the patient was not convalescing as expected. Fourteen of these 16 patients underwent a qualitative change in surgical approach.
In addition to low cardiac output state due to failure to achieve adequate physiology, deaths from ventricular failure, sudden bradycardiac arrest, sepsis, pulmonary hypertension, and specific procedural complications also occurred with sufficient frequency to guide improvement strategies for preoperative, intraoperative, and postoperative management.
The primary limitation of this study is its retrospective nature and reliance on review of medical records to determine mode of death. Despite this limitation, our study provides some important insights into improvement strategies that might reduce mortality after congenital heart surgery. Specifically, if surgical strategies or anatomy that are likely to result in inadequate postoperative physiology despite achievement of an adequate technical result can be identified preoperatively, different management strategies can be employed. Strategies aimed at reducing ventricular failure are also important, but are likely to have less of an impact than similar interventions in adult patients, as most patients who died after congenital heart surgery in this series had nearly normal ventricular function.
Fortunately, the mortality rate for infants and children with congenital heart disease has decreased substantially over the past 30 years [8–15]. Boneva and colleagues [1] found that death from heart defects decreased by 39% from 1979 to 1997, with the greatest decline in the infant population. Some authors report that surgery for infants weighing less than 2 kg can achieve a similar outcome as that for larger infants [8]. Ongoing assessment of modes of death may prove to be a useful way to guide future improvement efforts and further reduce risk in these patients.
|
References
|
|---|
- Boneva RS, Botto LD, Moore CA, Yang Q, Correa A, Erickson JD. Mortality associated with congenital heart defects in the United States: trends and racial disparities, 1979-1997 Circulation 2001;103:2376-2381.[Abstract/Free Full Text]
- Walker RE, Gauvreau K, Jenkins KJ. Health-related quality of life in children attending a cardiology clinic Pediatr Cardiol 2004;25:40-48.[Medline]
- Wypij D, Newburger JW, Rappaport LA. The effect of duration of deep hypothermic circulatory arrest in infant heart surgery on late neurodevelopment: the Boston Circulatory Arrest Trial J Thorac Cardiovasc Surg 2003;125:1397-1403.
- Gruber EM, Jonas RA, Newburger JW, Zurakowski D, Hansen DD, Laussen PC. The effect of hematocrit on cerebral blood flow velocity in neonates and infants undergoing deep hypothermic cardiopulmonary bypass Anesth Analg 1999;89:322-327.[Abstract/Free Full Text]
- O’Connor GT, Birkmeyer JD, Dacey LJ, et al. Results of a regional study of modes of death associated with coronary artery bypass grafting Ann Thorac Surg 1998;66:1323-1328.[Abstract/Free Full Text]
- Parr GV, Blackston EH, Kirklin JW. Cardiac performance and mortality early after intracardiac surgery in infants and young children Circulation 1975;51:867-875.[Abstract/Free Full Text]
- Guleserian KJ, Armsby LB, Thiagarajan RR, del Nido PJ, Mayer JE. Natural history of PA/IVS and RV-dependent coronary circulation managed by the single ventricle approach Ann Thorac Surg 2006;81:2250-2258.[Abstract/Free Full Text]
- Rossi AF, Seiden HW, Sadeghi AM, et al. The outcome of cardiac operations in infants weighing two kilograms or less J Thorac Cardiovasc Surg 1998;116:28-35.[Abstract/Free Full Text]
- Oechslin EN, Harrison DA, Connelly MS, Webb GD, Siu SC. Mode of death in adults with congenital heart disease Am J Cardiol 2000;86:1111-1116.[Medline]
- Dees E, Lin H, Cotton RB, Graham TP, Dodd DA. Outcome of preterm infants with congenital heart disease J Pediatr 2000;137:653-659.[Medline]
- Stark JF, Gallivan S, Davis K, et al. Assessment of mortality rates for congenital heart defects and surgeons performance Ann Thorac Surg 2001;72:169-175.[Abstract/Free Full Text]
- Jenkins KJ, Newburger JW, Lock JE, Davis RB, Coffman GA, Iezzoni LL. In-hospital mortality for surgical repair of congenital heart defects: preliminary observations of variants by hospital caseload Pediatrics 1995;95;:323-330.[Abstract/Free Full Text]
- Allen SW, Gauvreau K, Bloom BT, Jenkins KJ. Evidence based referral results in significantly reduced mortality after congenital heart surgery Pediatrics 2003;112:24-28.[Abstract/Free Full Text]
- Hannan EL, Racz M, Kavey RE, Quaegebeur JM, Williams R. Pediatric cardiac surgery: the effect of hospital and surgeon volume on in-hospital mortality Pediatrics 1998;101:963-969.[Abstract/Free Full Text]
- Birkmeyer JD, Siewers AE, Finalyson EVA, et al. Hospital volume and surgical mortality in the United States N Engl J Med 2002;346:1128-1137.[Abstract/Free Full Text]
Top
Abstract
Introduction
