If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Outcomes after first-stage palliation of single-ventricle heart disease are influenced by many factors, including the presence of residual lesions requiring reintervention. However, there is a dearth of information regarding the optimal timing of reintervention. We assessed if earlier reintervention would be favorably associated with in-hospital outcomes among patients requiring unplanned reinterventions after the Norwood operation.
Methods
This was a single-center, retrospective review of all patients who underwent the Norwood procedure from January 1997 to November 2017 and required a predischarge unplanned surgical or transcatheter reintervention on 1 or more subcomponent areas repaired at the index operation. Outcomes of interest included in-hospital mortality or transplant, postoperative hospital length of stay, and inpatient cost. Associations between timing of reintervention and outcomes were assessed using logistic regression (mortality or transplant) or generalized linear models (postoperative hospital length of stay and cost), adjusting for baseline patient-related and procedural factors.
Results
Of 500 patients who underwent the Norwood operation, 92 (18.4%) required an unplanned reintervention. Median time to reintervention was 12 days (interquartile range, 5-35 days). There were 31 (33.7%) deaths or transplants, median postoperative hospital length of stay was 49 days (interquartile range, 32-87 days), and median cost was $328,000 (interquartile range, $204,000-$464,000). On multivariable analysis, each 5-day increase in time to reintervention increased the odds of mortality or transplant by 20% (odds ratio, 1.2; 95% confidence interval, 1.1-1.3; P = .004). Longer time to reintervention was also significantly associated with greater postoperative hospital length of stay (P < .001) and higher cost (P < .001).
Conclusions
For patients requiring predischarge unplanned reinterventions after the Norwood operation, earlier reintervention is associated with improved in-hospital transplant-free survival and resource use.
Among patients requiring unplanned reinterventions for residual lesions after the Norwood operation, earlier reintervention is associated with improved in-hospital outcomes.
Neonates requiring unplanned reinterventions for residual lesions after the Norwood operation constitute an especially high-risk cohort. Earlier reintervention is associated with a reduced risk of in-hospital mortality or transplant, shorter length of stay, and lower inpatient cost compared with later reintervention. When feasible and clinically indicated, prompt reintervention should be considered.
See Commentary on page 447.
Hypoplastic left heart syndrome (HLHS) and other related single-ventricle lesions, constituting 2% to 4% of all congenital heart disease, are genetically heterogenous disorders that account for more than 20% of cardiac deaths in the first week of life if left unrepaired.
Despite advances in diagnosis, surgical techniques, and perioperative care, outcomes after first-stage palliation of single-ventricle heart disease (the Norwood operation) remain suboptimal. Several risk factors for adverse postoperative outcomes have been identified, including low birth weight, prematurity, and the presence of residual lesions requiring unplanned reinterventions before second-stage palliation.
among others. Recent evidence, including data from the Single Ventricle Reconstruction trial, has shown that the need for unplanned reinterventions before discharge from the Norwood operation is associated with longer time to first extubation, reduced in-hospital transplant-free survival (before discharge from first-stage palliation), increased risk of postdischarge reinterventions, and neurodevelopmental delay in infancy and childhood.
With a reported prevalence of up to 50% in national registries, unplanned reinterventions therefore impose a significant burden on patients and health systems.
Although residual lesions requiring reinterventions are a harbinger of adverse outcomes after the Norwood operation, the optimal timing of reintervention remains incompletely characterized. Moreover, it remains unclear if reinterventions that occur sufficiently early in the postoperative period, especially when the risk of mortality and morbidity is greatest, confer a survival advantage over the course of a child's first-stage hospitalization.
Therefore, we sought to assess if earlier reintervention is favorably associated with important in-hospital outcomes among patients requiring predischarge reinterventions after the Norwood operation.
Patients and Methods
Patient Population
Clinical and echocardiographic data from consecutive patients who underwent the Norwood procedure (the index operation) at a quaternary care center between January 1, 1997, and November 1, 2017, were retrospectively reviewed after Boston Children's Hospital Institutional Review Board approval (IRB-P00041313; Date of Approval/Exemption: January 20, 2022) and waiver of consent. Patients were included if they required a predischarge (in-hospital) unplanned surgical or transcatheter reintervention related to residual lesions involving 1 or more subcomponent areas repaired at the index surgery, including the atrial septum, neoaortic valve, modified BT shunt or RV-PA conduit, coronary arteries or Stansel anastomosis, proximal and distal aortic arch, and the pulmonary arteries (PAs) (catheter-based balloon dilatation or stenting of the branch PAs or surgical PA-plasty involving at least the branch PAs). Staged procedures for single-ventricle palliation and mediastinal reexploration for bleeding were not counted as reinterventions. We also excluded institution of extracorporeal membrane oxygenation (ECMO) as a reintervention because we specifically sought to assess the impact of time to unplanned reintervention for Norwood subcomponent-specific residua on postoperative outcomes.
mechanical circulatory support may be required for a variety of reasons unrelated to the aforementioned subcomponent-specific residual lesions, including low cardiac output syndrome and cardiovascular collapse.
Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation.
Clinical and Echocardiographic Indications for Reintervention
For all patients meeting entry criteria, the primary clinical indication that led to a reintervention or prompted further diagnostic workup that directly resulted in the specified reintervention was broadly categorized as persistent hypoxia or hypoxemia, acute desaturation, hemodynamic instability, or cardiac arrest. Furthermore, the echocardiographic severity of residual lesions before any unplanned reintervention was classified as mild or moderate to severe based on previously published criteria.
Specifically, patients with “mild” lesions had 1 of the following echocardiographic findings: (1) mild proximal arch stenosis with peak gradient 20 to 40 mm Hg or less than 30% narrowing by color Doppler jet width; (2) mild distal arch stenosis with peak gradient 20 to 40 mm Hg or less than 30% narrowing by color Doppler jet width; (3) mild coronary obstruction; (4) mild obstruction across the atrial septum with mean gradient 3 to 4 mm Hg; (5) mild neo-aortic valve regurgitation; (6) partial shunt or conduit obstruction or occlusion; or (7) mild branch PA stenosis. Likewise, patients with “moderate to severe” lesions met 1 of the following criteria: (1) moderate-severe proximal arch stenosis with peak gradient greater than 40 mm Hg or greater than 30% narrowing by color Doppler jet width; (2) moderate-severe distal arch stenosis with peak gradient greater than 40 mm Hg or greater than 30% narrowing by color Doppler jet width; (3) moderate-severe coronary obstruction; (4) moderate-severe obstruction across the atrial septum with mean gradient greater than 4 mm Hg; (5) moderate-severe neo-aortic valve regurgitation; (6) complete shunt or conduit obstruction or occlusion; or (7) moderate-severe branch PA stenosis.
Outcomes, Predictors, and Covariates
The primary of outcome was in-hospital (early) mortality or transplant. Other outcomes of interest included postoperative hospital length of stay (PHLOS) and inpatient cost of hospitalization. PHLOS was defined as the number of days from the Norwood operation to discharge. Of note, there were 4 patients (0.8%) who underwent an unplanned reintervention and remained in the hospital between the Norwood operation and second-stage palliation for a variety of reasons, including high-risk physiology, sociodemographic factors (eg, lack of appropriate home-monitoring facilities), and cumulative accrual of various minor and major complications. These children were not included in the analysis of PHLOS because their true length of stay could not be determined.
Furthermore, given the negligible proportion of these children relative to the overall study population (<1%), they were similarly excluded from the analysis of in-hospital mortality and inpatient cost to maintain uniformity in patient population. Inpatient costs were queried from an internal financial database that ties hospital account records with billing information specific to each encounter or patient visit. Costs were related only to inpatient hospital services, including intensive care unit and operating room use, medical supplies, laboratory tests and point-of-care evaluations, pharmacy-related services, radiographic and imaging services, and equipment, among others. Costs and fees related to professional services were excluded. All inpatient costs were inflated to the 2021 US dollar using the medical component of the consumer price index.
The primary predictor was time to reintervention, defined as the number of days from the Norwood operation to the first unplanned reintervention. In addition to timing of reintervention, several patient- and procedure-related covariates were assessed, including age, preterm birth (<37 gestational weeks), presence of at least 1 major preoperative risk factor, and a composite of various procedure-specific risk factors. System-specific preoperative risk factors included cardiopulmonary resuscitation, shock, mechanical circulatory support, mechanical ventilation, renal or hepatic failure, stroke, sepsis, and necrotizing enterocolitis. Procedural risk factors included ascending aortic diameter less than 2 mm, at least moderate left or right atrioventricular valve (AVV) regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructed pulmonary venous return.
Statistical Analysis
Summary statistics were reported as frequency (percentage) for categorical variables and median (interquartile range [IQR]) for continuous variables. The association between timing of reintervention and early mortality or transplant was evaluated using logistic regression. Model discrimination was assessed using the area under the receiver operating characteristic curve (C-statistic). The relationship between timing of reintervention and both PHLOS and inpatient cost of hospitalization was examined using generalized linear models, assuming a gamma distribution with a logarithmic link. Goodness-of-fit for all generalized linear models was determined using the coefficient of determination (R2). Multivariable models adjusted for all covariates chosen a priori. Odds ratios (ORs) and regression coefficients were reported with 95% confidence intervals (CIs). Stata version 15 (StataCorp LLC) was used throughout.
Results
Of 500 patients who underwent the Norwood operation during the study timeframe, 92 (18.4%) required at least 1 predischarge unplanned reintervention. HLHS was the primary diagnosis in 80 patients (87.0%); other primary defects included double-inlet left ventricle, double-outlet right ventricle with or without d-transposition of the great arteries, tricuspid atresia, and unbalanced atrioventricular canal (Table 1). A BT or modified BT shunt was used in 49 patients (53.3%); the remainder received an RV-PA conduit. Of note, postoperatively, 17 patients (18.5%) developed moderate or greater AVV regurgitation before discharge, death, or transplantation. Surgical AVV repair was performed in 7 patients (7.6%), but never in isolation (ie, always in conjunction with an unplanned reintervention for a residual lesion).
Table 1Baseline patient and procedural characteristics
Variable
Entire cohort (N = 92)
Patient characteristics
Age at Norwood operation (d)
4 (3-6)
Female sex
30 (32.6)
Weight at Norwood operation (kg)
3.2 (2.8-3.5)
Weight <2.5 kg
10 (10.9)
Preterm birth (<37 gestational wk)
15 (16.3)
Noncardiac anomaly, syndrome, or genetic abnormality
17 (18.5)
Primary diagnosis
HLHS
80 (87.0)
Double-outlet right ventricle
2 (2.2)
Double-outlet right ventricle, d-Transposition of the great arteries
2 (2.2)
Double-inlet left ventricle
4 (4.4)
Unbalanced atrioventricular canal
2 (2.2)
Tricuspid atresia
2 (2.2)
System-specific preoperative risk factor
Mechanical ventilation
25 (27.2)
Cardiopulmonary resuscitation or mechanical circulatory support
0 (0)
Renal failure
2 (2.2)
Stroke
2 (2.2)
Sepsis
3 (3.3)
Seizures, hepatic failure, necrotizing enterocolitis, or shock
8 (8.7)
At least 1 system-specific risk factor
29 (31.5)
Procedure-specific risk factor
Ascending aortic diameter <2.5 mm
15 (16.3)
At least moderate AVV regurgitation
6 (6.5)
Aortic atresia
48 (52.2)
Intact or restrictive atrial septum or obstructed pulmonary venous return
Of the 18 patients who required a second CPB run, the indications were as follows: revision of neo-aortic valve for at least mild regurgitation, 8 (44.4%); revision of atrial septectomy for mean gradient 5 mm Hg, 1 (5.6%); revision of Stansel anastomosis, 3 (16.7%); BT shunt revision, 2 (11.1%); arch revision for at least mild stenosis with peak gradient 20 to 50 mm Hg, 3 (16.7%); and global ventricular dysfunction, 1 (5.6%).
Based on the predischarge echocardiogram or echocardiogram immediately before in-hospital death or transplant.
None
7 (7.6%)
Trace or trivial
21 (22.8%)
Mild
47 (51.1%)
Moderate
16 (17.4%)
Severe
1 (1.1%)
Surgical era
1997-2003
23 (25.0)
2004-2010
33 (35.9)
2011-2017
36 (39.1)
Baseline patient-related and procedural characteristics are shown as median (IQR) or frequency (percentage). HLHS, Hypoplastic left heart syndrome; AVV, atrioventricular valve; BT, Blalock-Taussig; RV-PA, right ventricle-pulmonary artery; ECMO, extracorporeal membrane oxygenation.
∗ Of the 18 patients who required a second CPB run, the indications were as follows: revision of neo-aortic valve for at least mild regurgitation, 8 (44.4%); revision of atrial septectomy for mean gradient 5 mm Hg, 1 (5.6%); revision of Stansel anastomosis, 3 (16.7%); BT shunt revision, 2 (11.1%); arch revision for at least mild stenosis with peak gradient 20 to 50 mm Hg, 3 (16.7%); and global ventricular dysfunction, 1 (5.6%).
† Based on the predischarge echocardiogram or echocardiogram immediately before in-hospital death or transplant.
Patients requiring reinterventions involving the various subcomponent areas of the Norwood operation were as follows: proximal arch, 11 (12.0%); distal arch, 13 (14.1%); coronary arteries or Stansel anastomosis, 12 (13.0%); atrial septum, 8 (8.7%); neoaortic valve, 5 (5.4%); BT shunt or RV-PA conduit, 48 (52.2%); and branch PA, 25 (27.2%). The median time to reintervention in the entire cohort was 12 days (IQR, 5-35 days). The modality, clinical and echocardiographic indications, and timing of all subcomponent-specific reinterventions are summarized in Table 2. Overall, 11 patients (12.0%) received more than 1 surgical reintervention, 8 patients (8.7%) underwent more than 1 catheter-based reintervention, and 19 patients (20.7%) required a reintervention in more than 1 anatomic subcomponent of the Norwood operation. The most common anatomic area that required a reintervention was the BT shunt or RV-PA conduit. Among these 48 patients, 22 (45.8%) had a BT shunt. Indications for shunt reintervention included partial or complete shunt thrombosis or occlusion (related to elevated pulmonary vascular resistance or technical narrowing of the shunt itself) and reduced pulmonary blood flow due to diastolic run-off or a reactive pulmonary vascular bed. RV-PA conduit reinterventions (n/N 26/48; 54.2%) were performed solely for conduit stenosis. Nonparametric comparisons of source of pulmonary blood flow and in-hospital mortality or transplant (P = .83), PHLOS (P = .97), and inpatient cost (P = .10) did not yield any significant associations.
Table 2Subcomponent-specific reinterventions by modality, clinical indication, echocardiographic criteria, and timing
Refers to one of the following echocardiographic findings: (i) mild proximal arch stenosis with peak gradient (PG) 20-40 mm Hg or <30% narrowing by color Doppler jet width, (ii) mild distal arch stenosis with PG 20-40 mm Hg or <30% narrowing by color Doppler jet width, (iii) mild coronary obstruction, (iv) mild obstruction across the atrial septum with mean gradient 3-4 mm Hg, (v) mild neo-aortic valve regurgitation, (vi) partial shunt or conduit obstruction or occlusion, or (vii) mild branch PA stenosis.
Refers to one of the following echocardiographic findings: (i) moderate-severe proximal arch stenosis with PG > 40 mm Hg or >30% narrowing by color Doppler jet width, (ii) moderate-severe distal arch stenosis with PG > 40 mm Hg or >30% narrowing by color Doppler jet width, (iii) moderate-severe coronary obstruction, (iv) moderate-severe obstruction across the atrial septum with mean gradient >4 mm Hg, (v) moderate-severe neo-aortic valve regurgitation, (vi) complete shunt or conduit obstruction or occlusion, or (vii) moderate-severe branch PA stenosis.
Number of days from the Norwood operation to the first in-hospital reintervention.
3 (1-17)
42 (12-56)
4 (2-20)
37 (7-46)
9 (6-16)
7 (3-17)
24 (14-35)
Values are frequency (column percentage) or median (IQR). DKS, Damus-Kaye-Stansel anastomosis; BT, Blalock-Taussig; RV-PA, right ventricle-pulmonary artery; PA, pulmonary artery.
∗ Primary clinical indication that led to a reintervention or prompted further diagnostic workup that directly resulted in the specified reintervention.
† Residual lesion related directly to the reintervention under consideration.
‡ Refers to one of the following echocardiographic findings: (i) mild proximal arch stenosis with peak gradient (PG) 20-40 mm Hg or <30% narrowing by color Doppler jet width, (ii) mild distal arch stenosis with PG 20-40 mm Hg or <30% narrowing by color Doppler jet width, (iii) mild coronary obstruction, (iv) mild obstruction across the atrial septum with mean gradient 3-4 mm Hg, (v) mild neo-aortic valve regurgitation, (vi) partial shunt or conduit obstruction or occlusion, or (vii) mild branch PA stenosis.
§ Refers to one of the following echocardiographic findings: (i) moderate-severe proximal arch stenosis with PG > 40 mm Hg or >30% narrowing by color Doppler jet width, (ii) moderate-severe distal arch stenosis with PG > 40 mm Hg or >30% narrowing by color Doppler jet width, (iii) moderate-severe coronary obstruction, (iv) moderate-severe obstruction across the atrial septum with mean gradient >4 mm Hg, (v) moderate-severe neo-aortic valve regurgitation, (vi) complete shunt or conduit obstruction or occlusion, or (vii) moderate-severe branch PA stenosis.
‖ Number of days from the Norwood operation to the first in-hospital reintervention.
A total of 31 (33.7%) in-hospital deaths or transplants were observed. On univariable logistic regression, each 5-day increase in the time to reintervention increased the odds of early mortality or transplant by 20% (OR, 1.2, 95% CI, 1.1-1.3, P = .001). On multivariable analysis, time to reintervention (OR for every 5-day increase: 1.2, 95% CI, 1.1-1.3, P = .004) and the presence of at least 1 major systemic-specific preoperative risk factor (OR, 3.3, 95% CI, 1.1-9.8, P = .037) were significantly associated with early mortality or transplant (Table 3). Prematurity and a composite of various procedural risk factors were also associated with an increased risk of the primary outcome, but these risk estimates did not reach statistical significance. The discrimination of the timing of reintervention-only model, covariates-only model, and full multivariable model were 0.701, 0.748, and 0.801, respectively. The predicted probability of in-hospital mortality or transplant by time to reintervention, based on estimates from the full multivariable model, is shown in Figure 1.
Table 3Univariable and multivariable logistic regression models of in-hospital mortality or transplant
Odds for each 5-day increase in time to reintervention defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
2.6 (0.8-8.6)
.11
3.0 (0.8-11.6)
.12
Univariable and multivariable predictors of early mortality or transplant are shown here, with all covariates chosen a priori. Model C-statistic: time to reintervention only, 0.701; covariates only, 0.748; time to reintervention and covariates, 0.801. CI, Confidence interval.
∗ Odds for each 5-day increase in time to reintervention defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
† Odds for each 1-day increase in age.
‡ Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
§ Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
Figure 1Probability of in-hospital mortality or transplant by timing of reintervention. The predicted probability of in-hospital death or transplant by timing of reintervention is depicted. Time to reintervention was defined as the number of days from the Norwood operation to the first, unplanned, in-hospital reintervention for a residual lesion. All probability estimates were obtained from the corresponding multivariable logistic regression model of in-hospital mortality or transplant with time to reintervention as the primary predictor. The shaded area represents the 95% CI.
Overall median PHLOS was 49 days (IQR, 32-87 days), and median inpatient cost of hospitalization (rounded to the nearest thousand US dollars) was $328,000 (IQR, $204,000-$464,000). The distribution of PHLOS and inpatient cost by time to reintervention is shown in Figure 2. On both univariable and multivariable analyses, greater time to reintervention was significantly associated with longer PHLOS (both P < .001) (Table 4). The multivariable generalized linear model of PHLOS also identified older age and presence of at least 1 major preoperative risk factor as predictors of longer PHLOS, but these findings were not significant at the 0.05 level. Longer time to reintervention was also associated with higher cost on unadjusted (coefficient 0.057, 95% CI, 0.039-0.075, P < .001) and adjusted (coefficient 0.057, 95% CI, 0.037-0.075, P < .001) analyses (Table 5).
Figure 2Secondary outcomes of interest by timing of reintervention. A scatterplot of postoperative hospital length of stay (blue triangle) and inpatient cost of hospitalization (red circle) are displayed for the entire cohort. Time to reintervention was positively associated with both length of stay and inpatient cost. USD, US dollar.
Estimate for each 5-day increase in time to reintervention, defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
0.013 (–0.41 to 0.44)
.95
–0.11 (–0.49 to 0.28)
.59
Univariable and multivariable generalized linear models of postoperative hospital length of stay are shown here, assuming a gamma distribution with a logarithmic link, with all covariates chosen a priori. Model coefficient of determination: time to reintervention only, 0.46; covariates only, 0.062; time to reintervention and covariates, 0.50. CI, Confidence interval.
∗ Estimate for each 5-day increase in time to reintervention, defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
† Estimate for each 1-day increase in age.
‡ Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
§ Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
Estimate for each 5-day increase in time to reintervention, defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
0.30 (–0.047 to 0.66)
.089
0.23 (–0.033 to 0.50)
.086
Univariable and multivariable generalized linear models of inpatient cost of hospitalization are shown, assuming a gamma distribution with a logarithmic link, with all covariates chosen a priori. Model coefficient of determination: time to reintervention only, 0.65; covariates only, 0.074; time to reintervention and covariates, 0.66. CI, Confidence interval.
∗ Estimate for each 5-day increase in time to reintervention, defined as the number of days from the Norwood operation to the first in-hospital unplanned reintervention.
† Estimate for each 1-day increase in age.
‡ Presence of at least 1 major preoperative risk factor, including mechanical ventilation, shock, cardiopulmonary resuscitation, ECMO, renal failure, liver failure, stroke, seizure, sepsis, and necrotizing enterocolitis.
§ Presence of at least 1 procedural risk factor, including ascending aortic diameter less than 2 mm, at least moderate left or right AVV regurgitation, aortic atresia, and intact or restrictive atrial septum or obstructive pulmonary venous return.
The present study represents our 20-year experience with predischarge unplanned reinterventions after the Norwood operation. We found that earlier reintervention was significantly associated with a decreased risk of in-hospital mortality or transplant, shorter postoperative hospital length of stay, and lower inpatient cost of hospitalization, compared with later reintervention (Figure 3 and Video 1).
Figure 3Timing of unplanned predischarge reintervention in the Norwood population. Among patients with single-ventricle heart disease requiring in-hospital unplanned reinterventions for residual lesions after the Norwood operation, earlier reintervention is significantly associated with a decreased risk of early mortality or transplant, shorter postoperative hospital length of stay, and lower inpatient cost of hospitalization, compared with later reintervention. USD, US dollar; BT, Blalock-Taussig; RV-PA, right ventricle-pulmonary artery.
The need for an unplanned reintervention after neonatal cardiac surgery often portends adverse outcomes, including an increased risk of in-hospital mortality.
Unsurprisingly, this risk is exacerbated in patients after the Norwood procedure, which carries one of the highest mortality rates among all congenital cardiac operations.
Contemporary series have reported in-hospital mortality rates of up to 70% for patients requiring reinterventions after first-stage palliation, and our results compare favorably with these estimates.
Notably, within our study cohort, the proportion of in-hospital reinterventions increased over time, with the greatest number of reinterventions observed in the most recent surgical era. Although the underlying reasons remain nebulous, possible explanations include the relatively greater variety of contemporary transcatheter options available to treat various subcomponent-specific residual lesions, an increase in the number and level of complex operations in recent years, and evolving trends in the threshold for reintervention.
Of note, although AVV repair was required in less than 10% of patients postoperatively, AVV repair was not categorized as a reintervention per se, given that the AVV is not one of the index subcomponents of the Norwood procedure. When required in the postoperative period, always in conjunction with an unplanned reintervention as defined within the context of this study, the technique used was individualized to the patient's anatomy based on the location of regurgitation, annular diameter, and presence of leaflet prolapse or tethering. Surgical management strategies primarily involved partial annuloplasty, cleft closure, or commissuroplasty, and less commonly used chordal shortening and papillary muscle advancement.
The impact of timing of reintervention on outcomes after the Norwood operation has not been consistently reported or well studied in the literature. In a contemporary single-center series involving 100 patients with HLHS who underwent first-stage palliation using a RV-PA conduit, Doppler flow reversal within the RV-PA conduit to the right ventricle increased steadily between post-Norwood day 2 and discharge from first-stage palliation. However, the timing of conduit reintervention or its association with survival outcomes was not described.
In the present study, we demonstrated that the odds of mortality or transplant increased by approximately 20% for each 5-day increase in time to reintervention. The mechanisms underlying this observation are not entirely apparent, but likely relate to the timely identification of hemodynamically significant residual lesions with prompt treatment.
Our findings can only be generalized to patients who survive to a predischarge reintervention. It is conceivable that earlier reintervention leads to higher transplant-free survival because acutely sick neonates with major residual lesions die before a reintervention can be attempted. Alternatively, the increased risk of mortality seen with delayed reintervention may be attributed to the fact that residual lesions related to specific anatomic subcomponents of the Norwood operation warranting earlier reintervention have different hemodynamic consequences than residua necessitating later reintervention. As seen in our experience, some residual lesions, especially distal arch and branch PA residua, manifest only after the initial postoperative acute phase, whereas the risk of reintervention for residua related to coronary or shunt obstruction likely persists throughout the entire duration of the post-Norwood hospitalization.
Data from the Single Ventricle Reconstruction trial demonstrated that re-coarctation was not consistently predictable early in the post-Norwood course.
Late reintervention is often inevitable in instances where residua present closer to discharge from first-stage palliation.
In addition to higher transplant-free survival, patients undergoing earlier reintervention also demonstrated significantly lower resource use (shorter duration of hospitalization and decreased cost). These findings expand on prior work from our institution that revealed patients with single ventricles with major residual lesions or those requiring unplanned reinterventions after the Norwood operation confer a per-patient risk adjusted difference of $135,948 compared with patients with no residua.
Although it seems intuitive that a longer hospitalization leads to greater cost, data from national registries have shown that differences in PHLOS only account for approximately two-thirds of the cost variation not explained by post-Norwood complications.
It is plausible that later reintervention is associated with more complex physiology requiring a greater number of hospital resources. Of note, PHLOS is incorporated into various quality metrics in congenital heart surgery, including a composite morbidity metric proposed for the Congenital Heart Surgery Database.
However, when assessing outcomes after the Norwood operation, these metrics do not account for the longer length of stay inherently observed with patients requiring a later reintervention. Furthermore, longer PHLOS may itself be a risk factor for delayed reintervention. Enhancing our understanding of how specific residual lesions influence the timing of reintervention, as well as how timing impacts quality metrics such as PHLOS, is essential when evaluating national and center-specific variations in outcomes.
Given the importance of timely identification and treatment of residual lesions after the Norwood operation, a systematic method of early detection and prioritization of residua may be of benefit. Although the clinical indications for reintervention are diverse and vary by patient anatomy and physiology, we routinely use objective echocardiographic criteria for risk stratifying patients by the severity of residual lesions.
In turn, these criteria, based on the echocardiographic components of the Residual Lesion Score, may aid prognostication of high-risk patients and enable clinicians to determine the need for and optimal timing of reintervention. In prior studies, we have demonstrated that patients with severe residua involving various subcomponents of the Norwood operation have dramatically worse in-hospital and long-term outcomes compared with those with no or trivial residua.
Because residual lesions often evolve over time from the immediate postoperative period to discharge from first-stage palliation, with regard to both severity and anatomic modality,
we have recently instituted a system of standardized echocardiograms that are performed at regular, prespecified intervals in the post-Norwood period (eg, on postoperative days 3 and 9); these intervals are guided by the patient's clinical course and are determined in collaboration with the cardiac surgeon, cardiac intensivist, and cardiologist. These echocardiograms enable methodical ascertainment of the development of new residua and often prompt further imaging and invasive diagnostic evaluation. For instance, under this paradigm, a patient with borderline oxygenation with Residual Lesion Score Class 2 echocardiographic findings (mild distal aortic arch obstruction with peak gradient 25 mm Hg) will be identified as high risk for progression. Consequently, the interval until the next echocardiogram may be shortened or a preemptive completion catheterization may be performed to assess hemodynamics and suitability for reintervention. This protocol also allows for longitudinal analyses and may help providers to better understand the mechanisms that drive changes in residual lesions over time.
Study Limitations
This was a retrospective analysis with inherent bias related to missing data. Because this was a single-center study, the generalizability of our results may be limited. Furthermore, our findings may not be clinically relevant to patients who require reinterventions for residua that present solely toward the end of the Norwood hospitalization. Given that our study period spanned 2 decades, bias may have been introduced because of potential era effects secondary to evolving surgical practices and trends in perioperative management. Future endeavors will involve larger sample sizes and more granular assessment of subcomponent-specific reinterventions.
Conclusions
The present study constitutes a single institution's 20-year experience with first-stage palliation of single-ventricle heart disease and represents, to our knowledge, the first systematic evaluation of timing of predischarge unplanned reinterventions after the Norwood procedure. Specifically, for patients requiring reinterventions for residual lesions during the Norwood hospitalization, earlier reintervention is associated with a decreased risk of in-hospital mortality or transplant compared with later reintervention. Our findings may have implications for cost containment and suggest that the timing of reintervention can be a useful benchmark for resource use in the population with single-ventricle heart disease.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Important findings and implications: The main results of this study, along with their implications for clinical practice, are discussed. Specifically, among patients with single-ventricle heart disease requiring predischarge unplanned surgical or transcatheter reinterventions after the Norwood operation, the odds of in-hospital mortality or transplant increased by approximately 20% for each 5-day increase in time to reintervention. Time to reintervention was also positively associated with increased length of stay and inpatient cost of hospitalization. Our findings can only be generalized to patients who survive to a predischarge reintervention. Furthermore, the increased risk of mortality seen with delayed reintervention may be attributed to the fact that residual lesions related to specific anatomic components of the Norwood operation warranting earlier reintervention have different hemodynamic consequences than those necessitating later reintervention. Ultimately, a systematic protocol for identifying and treating residual lesions postoperatively is imperative to preempt the attendant adverse consequences of delayed reintervention. Video available at: https://www.jtcvs.org/article/S0022-5223(22)00517-7/fulltext.
Important findings and implications: The main results of this study, along with their implications for clinical practice, are discussed. Specifically, among patients with single-ventricle heart disease requiring predischarge unplanned surgical or transcatheter reinterventions after the Norwood operation, the odds of in-hospital mortality or transplant increased by approximately 20% for each 5-day increase in time to reintervention. Time to reintervention was also positively associated with increased length of stay and inpatient cost of hospitalization. Our findings can only be generalized to patients who survive to a predischarge reintervention. Furthermore, the increased risk of mortality seen with delayed reintervention may be attributed to the fact that residual lesions related to specific anatomic components of the Norwood operation warranting earlier reintervention have different hemodynamic consequences than those necessitating later reintervention. Ultimately, a systematic protocol for identifying and treating residual lesions postoperatively is imperative to preempt the attendant adverse consequences of delayed reintervention. Video available at: https://www.jtcvs.org/article/S0022-5223(22)00517-7/fulltext.
Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation.
Sengupta and colleagues1 from Boston Children's Hospital provide a data-driven argument for the proactive surveillance of patients who are not performing to expectations, demonstrating improved outcomes in those who have early correction of residual lesions. In the Norwood operation, 3 elements need to be satisfactory—the aortic arch, the Damus connection, and the source of pulmonary blood flow. In this series, 18% had 1 or more problems at a range of postoperative time points. Strikingly, the presence of a residual or new lesion escalated the risk of mortality or transplant to approximately 20%, whereas later identification was associated with a far greater risk.
00517-7&id=fx1.jpg){kind=link}
00517-7&id=fx2.jpg){kind=link}
00517-7&id=gr1.jpg){kind=link}
00517-7&id=gr2.jpg){kind=link}
00517-7&id=gr3.jpg){kind=link}
00517-7&id=fx4.jpg){kind=link}