Due to the improving care of children with congenital heart disease (CHD), the number of adult CHD (ACHD) patients is rapidly increasing,1 and most of these patients have a life expectancy well into adulthood.2 As they age, many of these patients will require further cardiac operations as adults, and previous studies have demonstrated low ACHD surgical mortality rates of 1.8% to 2.2%.3–5 Postcardiotomy extracorporeal life support (ECLS) has been well described in the pediatric congenital cardiac surgery population6–8 and adult general cardiac surgery population9,10 as a means to improve hospital survival. However, there are limited descriptions of the use of ECLS in the ACHD population with reported mortality rates of 54–67%.11,12 Additionally, predictors for mortality after ECLS in the ACHD population have not been described. Finally, among children, ECLS after Fontan surgery in single ventricle patients is associated with a high rate of mortality13 in comparison to ECLS after other pediatric heart surgeries.7
We sought to describe the characteristics and outcomes of ACHD patients supported with ECLS. We hypothesized that there would be identifiable demographic and clinical risk factors, including Fontan physiology, for mortality after ECLS use in this population and sought to evaluate this using a large, multicenter database.
Materials and Methods
Data were obtained from the Extracorporeal Life Support Organization (ELSO) registry, a multicenter international registry including clinical and administrative data from over 300 ECLS centers. Reporting centers submit data to the registry using a standardized collection form including demographic, clinical and procedural information, pre-ECLS and ECLS support characteristics, and ECLS complications. Extracorporeal Life Support Organization data use agreements allow the release of subsets of data for further analysis. Approval from the institutional review board at Seattle Children’s Hospital was obtained for these analyses.
Adult patients (age ≥ 18 years) who received ECLS after undergoing congenital heart surgery between January 1994 and October 2016 were identified using diagnostic and procedure codes. For inclusion into the study, we required an International Classification of Diseases-Ninth Edition, Clinical Modification (ICD-9-CM) diagnosis compatible with CHD ICD-9-CM codes (745.0–747.49). We excluded patients who had only one ELSO ICD-9-CM code for either aortic (745.32) or mitral (745.52) valve stenosis without an additional ICD-9-CM code for CHD as these patients may have isolated, acquired heart disease. Surgical complexity was stratified by Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery (STAT) mortality category using Current Procedural Terminology (CPT) codes,14,15 limiting procedures to only those that occurred before ECLS. Patients with a CHD diagnosis as above and with a CPT code of 33999 (unlisted procedure, cardiac surgery) were included as an “Unclassified” group. Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery category 5 was grouped with STAT category 4 due to the rarity of STAT category 5 procedures in adults, and STAT category 1 and 2 were grouped together due to similar outcomes between these groups. A separate variable denoting Fontan physiology was created for patients with a history of Fontan palliation (CPT codes 33615 or 33617) without history of a heart transplant (CPT code 33935 or 33945). Cardiovascular risk factors were identified using ICD-9-CM diagnostic codes. Extracorporeal Life Support Organization uniquely classifies the reason for ECLS as cardiac, pulmonary, or extracorporeal cardiopulmonary resuscitation (ECPR). We excluded all patients supported on ECLS for a pulmonary indication, limiting our analysis to those receiving cardiac and ECPR support.
Patient demographic information, diagnoses, procedures, pre-ECLS support, ECLS support, ECLS complications, and survival to hospital discharge were extracted for analysis. Patient age was dichotomized at the median age of 45 years into younger (18–44 years) versus older (45+years) groups. Weight was dichotomized due to a nonlinear relationship with mortality; a cut point of 100 kg was chosen by visual inspection of a Locally Weighted Scatterplot Smoothing plot of weight and mortality. Era was classified as 1994–2004, 2005–2008, 2009–2012, and 2013–2016, similar to Rood et al.13 Time from admission and intubation to ECLS cannulation were dichotomized with a cut point of 24 hours. Pre-ECLS support variables were classified similarly to Diddle et al.16 For patients with more than one ECLS run (n = 9), only the first run was included. Extracorporeal life support complications were categorized in a similar manner to Rood et al.,13 though excluding brain death from the neurologic complications due to the direct correlation between brain death and mortality and separating the respiratory and gastrointestinal complication groups into constituent components.
Continuous variables are reported as medians with interquartile range (IQR), whereas categorical and ordinal data are reported as numbers and percentages. Demographics, procedural, pre-ECLS support, and ECLS support data were compared between survivors and nonsurvivors, and a multivariate logistic regression model was constructed for mortality on ECLS. Nonparametric trends were used to evaluate change over time. A forward stepwise regression model was constructed, and variables with a p value less than 0.2 were candidates for inclusion, whereas variables were retained in the model if the adjusted p value was less than 0.05. Two separate models were constructed. The first model only contained variables that would be known before ECLS cannulation and identifies risk factors that clinicians can incorporate into pre-ECLS decision-making. The second model also includes postcannulation variables that impact ECLS survival and identifies complications that institutions should prioritize avoiding in ECLS patients. Given that patients who have had a heart transplant no longer have congenitally abnormal hearts, we performed a sensitivity analysis evaluating this model after exclusion of patients who underwent heart transplant. Based on the results of the multivariate mortality model, we created a multivariate logistic regression model evaluating predictors of neurologic complications in a manner similar to the mortality model. We also evaluated this model for all neurologic complications, including brain death as a sensitivity analysis. All models that include postcannulation variables are adjusted for time on ECLS using ECLS duration quartiles to adjust for varying time at risk of complications.
Continuous variables were compared using the Wilcoxon rank-sum test, and the χ2 and Fisher’s exact tests were used to examine categorical data as appropriate for categorical variables. Stata/SE (version 14.2; StataCorp LLC, College Station, TX) was used for data analysis.
Results
We identified 368 ACHD patients who received ECLS after cardiac surgery with an overall mortality rate of 61%. Overall mortality improved numerically over time from 74% to 56%, but this did not reach statistical significance in the overall cohort (p = 0.051). Mortality for younger ACHD patients (age 18–44 years) decreased from 75% between 1994 and 2004 to 51% between 2013 and 2016 (p = 0.02 for trend over time) (Figure 1, A). In contrast, mortality among older ACHD patients (age 45+ years) did not decrease over time (p = 0.66) (Figure 1, B). The number of ACHD patients supported with ECLS increased over time with a large rise in the final era; additionally, the proportion of older ACHD patients also increased over time (p = 0.01). Demographic variables were similar overall between survivors and nonsurvivors, although higher weight was more common among nonsurvivors (Table 1). Surgical complexity, as reflected by STAT categories, and underlying diagnoses were similar between survivors and nonsurvivors. However, Fontan physiology was significantly associated with mortality (89% mortality in patients with Fontan physiology vs. 59% mortality in those without, p = 0.002). Of the 27 patients with Fontan physiology, 12 had tricuspid atresia, 6 had common ventricle, and 1 had hypoplastic left heart syndrome, whereas the other 8 had other CHD diagnoses. A history of heart transplantation before ECLS was more common among survivors (p = 0.04).
Table 1. -
Demographic Characteristics of Adult Congenital Heart Surgery Patients Supported with Extracorporeal Life Support by Survivorship Status
| Characteristic |
Survivors (n = 142) |
Nonsurvivors (n = 226) |
p Value |
| Age (years) |
40 (26–57) |
45 (30–60) |
0.18 |
| Weight (kg)* |
67 (56–84) |
71 (58–91) |
0.12 |
| Weight groups |
|
|
0.01 |
| < 100 kg |
130 (92) |
182 (81) |
|
| ≥ 100 kg |
12 (8) |
42 (19) |
|
| Female* |
48 (34) |
91 (40) |
0.20 |
| Race* |
|
|
0.18 |
| White |
90 (64) |
143 (64) |
|
| Black |
4 (3) |
16 (7) |
|
| Other |
47 (33) |
65 (29) |
|
| Era |
|
|
0.051† |
| 1994–2004 |
13 (9) |
37 (16) |
|
| 2005–2008 |
27 (19) |
39 (17) |
|
| 2009–2012 |
28 (20) |
55 (24) |
|
| 2013–2016 |
74 (52) |
95 (42) |
|
| Indication |
|
|
0.42 |
| Cardiac |
130 (92) |
201 (89) |
|
| ECPR |
12 (8) |
25 (11) |
|
| Fontan physiology |
3 (2) |
24 (11) |
0.002 |
| History of heart transplant |
21 (15) |
18 (8) |
0.04 |
| Pulmonary hypertension |
7 (5) |
17 (8) |
0.33 |
| STAT category |
|
|
0.98 |
| 1–2 |
42 (30) |
70 (31) |
|
| 3 |
26 (18) |
39 (17) |
|
| 4–5 |
60 (42) |
93 (41) |
|
| Unclassified |
14 (10) |
24 (11) |
|
| Common diagnoses‡ |
| Left sided obstructive lesions |
27 (19) |
47 (21) |
0.68 |
| Aortic anomalies |
8 (6) |
16 (7) |
0.58 |
| Ventricular septal defect |
28 (20) |
50 (22) |
0.58 |
| Transposition of the great arteries |
14 (10) |
21 (9) |
0.86 |
| Tetralogy of Fallot |
15 (11) |
17 (8) |
0.31 |
| Double outlet right ventricle |
9 (6) |
12 (5) |
0.68 |
| Ebstein’s anomaly |
10 (7) |
17 (8) |
0.86 |
| Atrioventricular septal defect |
7 (5) |
10 (4) |
0.82 |
| Pulmonary valve disease |
8 (6) |
20 (9) |
0.26 |
| Cardiovascular risk factors |
| Coronary artery disease |
29 (20) |
41 (18) |
0.59 |
| Hyperlipidemia |
5 (4) |
8 (4) |
0.99 |
| Hypertension |
5 (4) |
16 (7) |
0.15 |
| Diabetes mellitus type 2 |
5 (4) |
9 (4) |
0.82 |
*p values exclude comparison of missing patients. Weight missing <1%, Gender missing <1%, Race missing <1%.
†p value for nonparametric trend. Percentages may not sum to 100% due to rounding.
‡Diagnoses are not mutually exclusive and previously repaired diagnoses are included.
ECPR, extracorporeal cardiopulmonary resuscitation; STAT, Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery.
;)
Figure 1.: A: Trends in utilization and hospital survival after extracorporeal life support among adult congenital heart surgery patients age 18–44 years. p = 0.02 for mortality trend by era. B: Trends in utilization and hospital survival after extracorporeal life support among adult congenital heart surgery patients age 45+ years. p = 0.66 for mortality trend by era. OR, odds ratio.
Pre-ECLS support and ECLS characteristics by survivor status are shown in Table 2. More nonsurvivors had a time from admission to cannulation > 24 hours, as well as a time from intubation to cannulation > 24 hours. pH on arterial blood gas and mean arterial blood pressure were similar between groups. Precannulation support was mostly similar between groups, but nonsurvivors were more likely to be supported with neuromuscular blockade than survivors. Extracorporeal life support characteristics are also shown in Table 2 and were similar between groups, including duration of ECLS support.
Table 2. -
Clinical and ECLS Characteristics of Survivors and Nonsurvivors Among Adult Congenital Heart Surgery Patients
| Characteristic |
Survivors (n = 142) |
Nonsurvivors (n = 226) |
p Value |
| Time to ECLS cannulation > 24 hr* |
69 (53) |
146 (69) |
0.003 |
| Time from intubation to cannulation > 24 hr* |
37 (27) |
83 (40) |
0.02 |
| pH on arterial blood gas* |
7.31 (7.23–7.37) |
7.28 (7.19–7.37) |
0.23 |
| Mean arterial blood pressure (mm Hg)* |
53 (42–62) |
55 (42–64) |
0.79 |
| Precannulation support |
|
|
|
| Bicarbonate |
18 (13) |
45 (20) |
0.07 |
| Vasodilators |
18 (13) |
26 (12) |
0.74 |
| Pulmonary hypertension medications |
18 (13) |
36 (16) |
0.39 |
| Neuromuscular blockers |
27 (19) |
70 (31) |
0.01 |
| Ventricular assist device |
3 (2) |
5 (2) |
0.95 |
| Intra-aortic balloon pump |
24 (17) |
54 (24) |
0.11 |
| Pre-ECLS arrest |
34 (24) |
58 (26) |
0.71 |
| Venoarterial ECLS |
139 (98) |
222 (98) |
0.82 |
| ECLS flow at 4 hr (L/min)* |
3.6 (2.8–4.3) |
3.8 (3.0–4.4) |
0.30 |
| ECLS flow at 24 hr (L/min)* |
3.8 (2.8–4.5) |
3.8 (2.9–4.5) |
0.93 |
| Duration of ECLS support (hr)* |
95 (57–140) |
101 (42–202) |
0.48 |
| Cannulation site* |
|
|
0.10 |
| Chest |
62 (50) |
113 (55) |
|
| Neck |
8 (6) |
24 (12) |
|
| Femoral |
55 (44) |
70 (34) |
|
Time to cannulation missing 7%, time from intubation to cannulation missing 6%, pH missing 23%, mean arterial blood pressure missing 42%, 4 hr pump flow missing: 7%, 24 hr pump flow missing: 17%, duration of ECLS support missing: 2%, cannulation site missing: 10%. Percentages may not sum to 100% due to rounding.
*p values exclude comparison of missing patients.
ECLS, extracorporeal life support.
Complications are shown in Table 3. Overall, 90% of patients experienced at least one complication, and nonsurvivors experienced a higher number of complications than survivors. Mechanical, neurologic, and renal complications were associated with mortality. Tamponade during ECLS and pH < 7.2 on ECLS were also associated with mortality.
Table 3. -
Complications in Adult Congenital Heart Disease Patients Supported with ECLS
| Characteristic |
Survivors (n = 142) |
Nonsurvivors (n = 226) |
p Value |
| No complications |
16 (11) |
20 (9) |
0.45 |
| Mechanical complications |
12 (9) |
43 (19) |
0.01 |
| Circuit related |
10 (7) |
25 (11) |
|
| Air in circuit |
0 (0) |
9 (4) |
|
| Cannula related |
2 (1) |
13 (6) |
|
| Circuit thrombus |
28 (20) |
43 (19) |
0.87 |
| Surgical bleeding |
57 (40) |
112 (50) |
0.08 |
| Neurologic injury |
6 (4) |
35 (16) |
0.001 |
| Seizures |
3 (2) |
14 (6) |
|
| Cerebral infarction |
2 (1) |
14 (6) |
|
| Intracranial hemorrhage |
2 (1) |
10 (4) |
|
| Renal failure |
52 (37) |
137 (61) |
<0.001 |
| Serum creatinine > 1.5 mg/dL |
24 (17) |
89 (39) |
|
| Renal replacement therapy |
41 (29) |
117 (52) |
|
| ECLS-related |
| Need for inotropes on ECLS |
89 (63) |
155 (69) |
0.24 |
| CPR on ECLS |
4 (3) |
17 (8) |
0.06 |
| pH < 7.2 on ECLS |
3 (2) |
25 (11) |
0.002 |
| Cardiac complications |
| Arrhythmia |
23 (16) |
52 (23) |
0.11 |
| Tamponade during ECLS |
7 (5) |
25 (11) |
0.04 |
| Respiratory complications |
| Pneumothorax |
2 (1) |
3 (1) |
0.95 |
| Pulmonary hemorrhage |
4 (3) |
17 (8) |
0.06 |
| Bloodstream infection |
29 (20) |
34 (15) |
0.18 |
| Metabolic complications |
| Hypoglycemia (blood glucose < 40 mg/dL) |
1 (1) |
5 (2) |
0.27 |
| Hyperglycemia (blood glucose > 240 mg/dL) |
19 (13) |
40 (18) |
0.27 |
| Gastrointestinal complications |
| Gastrointestinal hemorrhage |
4 (3) |
8 (4) |
0.70 |
| Hyperbilirubinemia |
14 (10) |
34 (15) |
0.15 |
| Limb complications |
5 (4) |
7 (3) |
0.82 |
| Total number of complications |
2 (1, 4) |
4 (2, 5) |
<0.001 |
CPR, cardiopulmonary resuscitation; ECLS, extracorporeal life support.
We created a multivariate logistic regression model examining mortality using pre-ECLS characteristics. In this model (Figure 2), Fontan physiology (odds ratio [OR]: 5.7; 95% CI: 1.6–20.0), weight over 100 kg (OR: 2.6; 95% CI: 1.3–5.4), female gender (OR: 1.6; 95% CI: 1.001–2.6), ECLS cannulation greater than 24 hours after admission (OR: 2.0; 95% CI: 1.2–3.2), and precannulation neuromuscular blockade (OR: 1.9; 95% CI: 1.1–3.3) were associated with mortality. After adding ECLS characteristics and ECLS duration to the model, renal complications (OR: 3.0; 95% CI: 1.7–5.2), neurologic complications (OR: 4.7; 95% CI: 1.5–15.2), and pulmonary hemorrhage (OR: 6.4; 95% CI: 1.3–32.3) were associated with mortality, whereas Fontan physiology was no longer associated with mortality (OR: 3.0; 95% CI: 0.8–11.2). A sensitivity analysis of the final model, excluding patients with a history of a heart transplant yielded similar results with the exception that weight over 100 kg was no longer associated with mortality. Although patients with a history of a heart transplant no longer have CHD, short-term post-transplant mortality of ACHD patients is worse than non-ACHD patients,17 and it is likely this is related to pretransplant anatomic and physiologic issues. For this reason, we retained the heart transplant patients in the primary analysis.
Figure 2.: Precannulation and postcannulation multivariate logistic regression models for mortality after extracorporeal life support in adult congenital heart surgery patients. Delayed cannulation: greater than 24 hours from admission to cannulation. OR, odds ratio.
We further explored the association between Fontan physiology and mortality through the iterations of the multivariate model (Figure 3). Fontan physiology was initially associated with mortality in the precannulation model. This association persisted after the addition of pulmonary hemorrhage as well as renal complications to the model. However, Fontan physiology was not associated with mortality after the addition of neurologic complications to the model.
Figure 3.: Iterative multivariate models for the association of Fontan physiology with mortality after extracorporeal life support in adult congenital heart surgery patients. OR, odds ratio.
We then evaluated predictors of neurologic complications (see Table, Supplemental Digital Content, https://links.lww.com/ASAIO/A483). Higher weight, ECPR, Fontan physiology, lower pH, pre-ECLS arrest, higher 24 hour ECLS flow, and renal failure were more common among patients with a neurologic complication; the total number of non-neurologic complications was slightly higher among patients with neurologic complications.
Finally, we constructed a multivariate model for experiencing a neurologic complication using the variables from the Table, Supplemental Digital Content, https://links.lww.com/ASAIO/A483 as candidate variables. In this model (Figure 4), only Fontan physiology (OR: 8.2; 95% CI: 3.3–20.9) and weight over 100 kg (OR: 3.1; 95% CI: 1.4–6.8) were associated with increased risk of a neurologic complication. A sensitivity analysis that included brain death in the definition of neurologic complications showed similar findings (data not shown).
Figure 4.: Multivariate logistic regression model for Neurologic complications after extracorporeal life support in adult congenital heart surgery patients. ECLS, Extracorporeal Life Support; OR, odds ratio.
Discussion
In this large, multicenter study, we identified precannulation risk factors for ECLS mortality after ACHD surgery including Fontan physiology, weight over 100 kg, female gender, prolonged time to ECLS cannulation, and preoperative neuromuscular blockade. Fontan physiology is also a risk factor for neurologic complications on ECLS. This suggests that neurologic complications may be a mediator for increased ELCS mortality among Fontan patients. These findings can inform medical decision-making among clinicians that care for ACHD patients.
Demographics
This study is the largest study of ECLS outcomes in ACHD surgery patients using validated, clinically oriented data. The use of ECLS among adult patients has been increasing over the last 10 years,18 and our data demonstrate a similar trend for ACHD patients with a large increase in the most recent era (46% of all included cases occurred between 2013 and 2016). Although the overall mortality (61%) found in this study was high, the mortality for younger patients in the final era (51%) approaches the mortality rate for pediatric patients placed on ECLS for cardiac (43%) and ECPR (57%) indications between 2009 and 2017.19 A previous study examining ECLS outcomes after postcardiotomy shock in adult patients found a mortality rate between 80% and 85% between the years of 2012 and 2016, in spite of a large increase in the volume of adult ECLS cases overall during this time.10 Additionally, a recent study on the use of ECLS in elderly patients between 2011 and 2015 did not identify a trend toward improved outcomes over time.20 Taken together, previous studies do not demonstrate an improvement in adult ECLS outcomes over time, despite the recent large increase in use. However, we demonstrate that ECLS outcomes are improving in a specific population of younger adults with ACHD. The reasons for this improved outcome in younger patients could not be evaluated in our study but may include increased familiarity with equipment, technological improvements, lower rates of comorbidities, possible treatment at pediatric versus adult ECLS centers, and better patient selection. Interestingly, although the proportion of older adult patients increased over time, this increased volume did not result in improved outcomes for these patients. Possible reasons for the lack of improvement in the older patient population include unmeasured comorbidities and difficulty identifying which patients are most likely to benefit from this therapy as well as the possibility that physiologic changes of aging may result in worse outcomes with the continuous flow circulation provided by ECLS.
Precannulation Model
We identified five risk factors for mortality in the precannulation model that can aid clinicians and families when making pre-ECLS decisions. First, Fontan physiology was significantly associated with mortality and was the pre-ECLS variable that had the greatest association with mortality. Given the unique characteristics of Fontan physiology including the lack of a subpulmonary ventricle, lack of an atrium to collect systemic venous drainage, reliance on passive venous return to the pulmonary vascular tree, and chronically elevated central venous pressure, it is not unexpected that Fontan physiology is a risk factor for ECLS mortality. This is similar to pediatric ECLS reports, which have shown a mortality rate of 65% in patients with Fontan physiology, higher than the overall ECLS mortality rate for CHD.7,13 Second, weight more than 100 kg was also a significant risk factor for mortality, similar to previous studies of adult ECLS21 and congenital heart surgery in older children and adults.22 Third, female gender was also a risk factor for mortality; this has previously been documented after pediatric cardiac surgery,23 ACHD surgery,24 and ECLS for acute myocarditis.25 Fourth, we found that ECLS cannulation more than 24 hours after admission was associated with mortality. This finding may reflect a negative impact of delayed institution of ECLS after surgery or maybe a marker of prolonged attempts at conventional support before ECLS. Alternatively, this may represent patients with complex comorbidities or illnesses that necessitated preoperative admission to the intensive care unit before surgery, who subsequently experienced a poor outcome. Although it is not possible to change preoperative issues leading to hospitalization before surgery, this data may support early institution of ECLS after surgery in patients with a concerning clinical status. Unfortunately, the ELSO dataset does not contain sufficient data to analyze the temporal relationship between surgery and ECLS initiation or to identify which patients were cannulated in the operating room. Finally, we found that precannulation neuromuscular blockade was associated with increased mortality. This finding is likely a surrogate of severe critical illness requiring decreased metabolic demand for treatment and is reflective of the degree of illness of the pre-ECLS population.26,27 Overall, the precannulation model provides information that can be used by clinicians when deciding whether to initiate ECLS after cardiac surgery in this patient population.
Postcannulation Model
The association of complications with ECLS mortality dominated the postcannulation model. In this model, weight more than 100 kg, female gender, and ECLS cannulation more than 24 hours after admission remained risk factors for mortality, but the other two variables in the precannulation model did not remain associated with mortality. The persistence of these factors suggests a direct relationship with mortality or one that is mediated through other mechanisms that are unmeasured in this analysis. Conversely, the effect of Fontan physiology and neuromuscular blockade on mortality may be mediated through complications as they are no longer associated with mortality after the addition of complications to the model. Female gender remained a risk factor for mortality in the postcannulation model, similar to previous reports of CHD surgery and ECLS.23–25 However, no clear reason for this finding has previously been identified, and the available data from our study also do not reveal an etiology. Possible reasons for this finding include differences in biology between genders, such as hormones, heart/great vessel size, and the prevalence of other risk factors as well as health system factors such as referral bias and differential care between genders. Pulmonary hemorrhage was also associated with mortality, as has previously been documented after ECPR.28 Neurologic complications have previously been found in 6–17% of postcardiotomy and adult ECLS patients9,29,30 and are associated with increased ECLS mortality.31 This association with ECLS mortality may be due to changes in neurologic status or radiological findings that necessitate rapid discontinuation of ECLS. Finally, we noted that renal complications were also a significant risk factor for mortality, consistent with previous reports.32,33 The postcannulation model identifies risk factors that modify the risk of mortality once a patient has been placed on ECLS. Although this does not affect the decision to initiate ECLS, it can guide clinicians regarding monitoring for complications and how these complications potentially affect the risk of mortality. Additionally, identifying complications that have the greatest association with mortality allows ECLS centers to identify “high-risk” complications that would be potential targets for ECLS quality improvement initiatives.
Neurologic Complication Model
Given the interrelation between Fontan physiology, mortality, and neurologic complications, we further explored the association between Fontan physiology and neurologic complications. Previous reports of ECLS mortality after Fontan surgery in children noted that neurologic injury, surgical bleeding, and renal failure were associated with increased mortality.13 In our analysis, Fontan physiology and weight more than 100 kg were the only predictors of experiencing an ECLS neurologic complication. The higher central venous pressures obligate in Fontan physiology may contribute to decreased end-organ resilience, including in the brain, both before and during ECLS, potentially predisposing patients with Fontan physiology to increased neurologic complications. We also noted that elevated weight was a risk factor for neurologic complications, which has not previously been described. Elevated weight and obesity are risk factors for increased atherosclerotic disease,34 which can manifest in the cerebrovascular system as well as the cardiovascular system. Additionally, patients with elevated weight may be more likely to experience difficulty with cannulation due to local adiposity. Finally, it may also be more difficult to manage the amount of ECLS flow needed for adequate neurologic perfusion in patients with elevated weight given the variation in cardiac output needs related to weight.
Role of Fontan Physiology
Fontan physiology allows many patients to survive to adulthood that would not have otherwise been able to do so, but it is fraught with multisystem issues related to chronically elevated central venous pressure (particularly including liver dysfunction), lack of a subpulmonary ventricle, and difficulty augmenting cardiac output in times of increased demand.35 Given previous reports of elevated ECLS mortality in children after the Fontan operation,13 it is reasonable that Fontan physiology would be a risk factor for mortality among adult patients, as it is in the precannulation model. However, during iterative additions of ECLS complications to the precannulation model (Figure 3), Fontan physiology was no longer associated with ECLS mortality after the addition of neurologic complications. Additionally, Fontan physiology itself is a strong risk factor for neurologic complications (Figure 4). Taken together, this suggests that the association between Fontan physiology and mortality may be mediated through increased complications, particularly neurologic complications. This suggests that efforts to identify and reduce neurologic complications on ECLS may have a disproportionate impact on survival for Fontan patients.
Limitations
This study has several important limitations, mostly related to the use of observational data. Although the ELSO registry is used extensively for outcomes research, the data it contains are voluntarily reported and are subject to inaccurate classification as well as missing data. In order to identify ACHD patients, we used all available diagnostic and procedure codes, cross-referencing these codes to obtain the highest-fidelity classification possible. However, this approach remains subject to inaccurate coding. Similarly, although we used strict rules to identify patients who had Fontan Physiology, we cannot completely ascertain that all patients with this current physiology were captured, and we were not able to classify these patients by specifics of their Fontan circulation including fenestration status and type of Fontan. We purposefully excluded patients with isolated mitral and aortic disease to minimize the inclusion of acquired heart disease; although this likely excluded some patients with congenital single valve disease, we felt that it was more important to exclude the patients with acquired heart disease than include these CHD patients. These single valve patients are also more likely to be evaluated in other studies of ECLS outcomes after valve surgery. We also did not have data to calculate severity of illness scores before cannulation or a comparison group of ACHD patients who were not placed on ECLS. Given the absence of height in the available data, we were not able to determine body mass index or obesity status. We also could not evaluate the role of center volume given data limitations. Additionally, we are unable to account for other clinical data such as anticoagulation and transfusion strategies. Finally, although this is the largest published study of ECLS outcomes in ACHD patients using a clinically oriented dataset, the total number of patients is still relatively low, limiting the number of comparisons that can be made and the conclusions that can be drawn.
Conclusions
This article identifies both precannulation and postcannulation risk factors for ECLS mortality after ACHD surgery. Patients with Fontan physiology, by facing higher risks of neurologic ECLS complications, are at increased risk of ECLS mortality. Given the rapid increase in ECLS use, understanding risk factors for ACHD patients receiving ECLS will aid clinicians in preoperative planning and counseling.
References
1. Marelli AJ, Ionescu-Ittu R, Mackie AS, Guo L, Dendukuri N, Kaouache M: Lifetime prevalence of congenital heart disease in the general population from 2000 to 2010. Circulation. 2014.130:749–756.
2. van der Bom T, Mulder BJ, Meijboom FJ, et al.: Contemporary survival of adults with congenital heart disease. Heart. 2015.101:1989–1995.
3. Fuller SM, He X, Jacobs JP, et al.: Estimating mortality risk for adult congenital heart surgery: An analysis of The Society of Thoracic Surgeons Congenital Heart Surgery Database. Ann Thorac Surg. 2015.100:1728–1735; discussion 1735.
4. Kim YY, Gauvreau K, Bacha EA, Landzberg MJ, Benavidez OJ: Risk factors for death after adult congenital heart surgery in pediatric hospitals. Circ Cardiovasc Qual Outcomes. 2011.4:433–439.
5. Mascio CE, Pasquali SK, Jacobs JP, Jacobs ML, Austin EH 3rd: Outcomes in adult congenital heart surgery: analysis of the Society of Thoracic Surgeons database. J Thorac Cardiovasc Surg. 2011.142:1090–1097.
6. Bratton SL, Chan T, Barrett CS, Wilkes J, Ibsen LM, Thiagarajan RR: Metrics to assess extracorporeal membrane oxygenation utilization in pediatric cardiac surgery programs. Pediatr Crit Care Med. 2017.18:779–786.
7. Mascio CE, Austin EH 3rd, Jacobs JP, et al.: Perioperative mechanical circulatory support in children: An analysis of the Society of Thoracic Surgeons Congenital Heart Surgery Database. J Thorac Cardiovasc Surg. 2014.147:658–664: discussion 664.
8. Misfeldt AM, Kirsch RE, Goldberg DJ, et al.: Outcomes of single-ventricle patients supported with extracorporeal membrane oxygenation. Pediatr Crit Care Med. 2016.17:194–202.
9. Rastan AJ, Dege A, Mohr M, et al.: Early and late outcomes of 517 consecutive adult patients treated with extracorporeal membrane oxygenation for refractory postcardiotomy cardiogenic shock. J Thorac Cardiovasc Surg. 2010.139:302–311, 311.e1.
10. Whitman GJ: Extracorporeal membrane oxygenation for the treatment of postcardiotomy shock. J Thorac Cardiovasc Surg. 2017.153:95–101.
11. Acheampong B, Johnson JN, Stulak JM, et al.: Postcardiotomy ECMO support after high-risk operations in adult congenital heart disease. Congenit Heart Dis. 2016.11:751–755.
12. Dolgner SJ, Krieger EV, Wilkes J, et al.: Predictors of extracorporeal membrane oxygenation support after surgery for adult congenital heart disease in children’s hospitals. Congenit Heart Dis. 2019.14:559–570.
13. Rood KL, Teele SA, Barrett CS, et al.: Extracorporeal membrane oxygenation support after the Fontan operation. J Thorac Cardiovasc Surg. 2011.142:504–510.
14. Chan T, Di Gennaro J, Farris RW, Radman M, McMullan DM: Racial and ethnic variation in pediatric cardiac extracorporeal life support survival. Crit Care Med. 2017.45:670–678.
15. O’Brien SM, Clarke DR, Jacobs JP, et al.: An empirically based tool for analyzing mortality associated with congenital heart surgery. J Thorac Cardiovasc Surg. 2009.138:1139–1153.
16. Diddle JW, Almodovar MC, Rajagopal SK, Rycus PT, Thiagarajan RR: Extracorporeal membrane oxygenation for the support of adults with acute myocarditis. Crit Care Med. 2015.43:1016–1025.
17. Davies RR, Russo MJ, Yang J, Quaegebeur JM, Mosca RS, Chen JM: Listing and transplanting adults with congenital heart disease. Circulation. 2011.123:759–767.
18. Thiagarajan RR, Barbaro RP, Rycus PT, et al.: Extracorporeal Life Support Organization Registry international report 2016. ASAIO J. 2017.63:60–67.
19. Barbaro RP, Paden ML, Guner YS, et al.: Pediatric Extracorporeal Life Support Organization registry international report 2016. ASAIO J. 2017.63:456–463.
20. Lorusso R, Gelsomino S, Parise O, et al.: Venoarterial extracorporeal membrane oxygenation for refractory cardiogenic shock in elderly patients: Trends in application and outcome from the Extracorporeal Life Support Organization (ELSO) Registry. Ann Thorac Surg. 2017.104:62–69.
21. Schmidt M, Burrell A, Roberts L, et al.: Predicting survival after ECMO for refractory cardiogenic shock: the survival after veno-arterial-ECMO (SAVE)-score. Eur Heart J. 2015.36:2246–2256.
22. O’Byrne ML, Kim S, Hornik CP, et al.: Effect of obesity and underweight status on perioperative outcomes of congenital heart operations in children, adolescents, and young adults: An analysis of data from the Society of Thoracic Surgeons Database. Circulation. 2017.136:704–718.
23. Seifert HA, Howard DL, Silber JH, Jobes DR: Female gender increases the risk of death during hospitalization for pediatric cardiac surgery. J Thorac Cardiovasc Surg. 2007.133:668–675.
24. Beurtheret S, Tutarel O, Diller GP, et al.: Contemporary cardiac surgery for adults with congenital heart disease. Heart. 2017.103:1194–1202.
25. Rajagopal SK, Almond CS, Laussen PC, Rycus PT, Wypij D, Thiagarajan RR: Extracorporeal membrane oxygenation for the support of infants, children, and young adults with acute myocarditis: A review of the Extracorporeal Life Support Organization registry. Crit Care Med. 2010.38:382–387.
26. deBacker J, Hart N, Fan E: Neuromuscular blockade in the 21
st Century management of the critically ill patient. Chest. 2017.151:697–706.
27. Murray MJ, Cowen J, DeBlock H, et al.; Task Force of the American College of Critical Care Medicine (ACCM) of the Society of Critical Care Medicine (SCCM), American Society of Health-System Pharmacists, American College of Chest Physicians: Clinical practice guidelines for sustained neuromuscular blockade in the adult critically ill patient. Crit Care Med. 2002.30:142–156.
28. Conrad SJ, Bridges BC, Kalra Y, Pietsch JB, Smith AH: Extracorporeal cardiopulmonary resuscitation among patients with structurally normal hearts. ASAIO J. 2017.63:781–786.
29. Lorusso R, Barili F, Mauro MD, et al.: In-hospital neurologic complications in adult patients undergoing venoarterial extracorporeal membrane oxygenation: results from the Extracorporeal Life Support Organization Registry. Crit Care Med. 2016.44:e964–e972.
30. Wu MY, Lin PJ, Lee MY, et al.: Using extracorporeal life support to resuscitate adult postcardiotomy cardiogenic shock: Treatment strategies and predictors of short-term and midterm survival. Resuscitation. 2010.81:1111–1116.
31. Sutter R, Tisljar K, Marsch S: Acute neurologic complications during extracorporeal membrane oxygenation: A systematic review. Crit Care Med. 2018.46:1506–1513.
32. Lee SW, Yu MY, Lee H, et al.: Risk factors for acute kidney injury and in-hospital mortality in patients receiving extracorporeal membrane oxygenation. PLoS One. 2015.10:e0140674
33. Yan X, Jia S, Meng X, et al.: Acute kidney injury in adult postcardiotomy patients with extracorporeal membrane oxygenation: evaluation of the RIFLE classification and the Acute Kidney Injury Network criteria. Eur J Cardiothorac Surg. 2010.37:334–338.
34. Hinnouho GM, Czernichow S, Dugravot A, et al.: Metabolically healthy obesity and the risk of cardiovascular disease and type 2 diabetes: The Whitehall II cohort study. Eur Heart J. 2015.36:551–559.
35. Rychik J: The relentless effects of the Fontan paradox. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2016.19:37–43.
