Search

Background: The Pediatric Interagency Registry for Mechanical Circulatory Support (Pedimacs) provides detailed understanding on pediatric patients supported with ventricular assist devices (VADs). We sought to identify important variables affecting the incidence of stroke in pediatric VADs., Methods: Between 2012 and 2022, 1463 devices in 1219 patients were reported to Pedimacs from 40 centers in patients aged <19 years at their first VAD implantation. Multiphase parametric hazard modeling was used to identify risk factors for stroke among all device types., Results: Of the 1219 patients, the most common devices were implantable continuous (472 [39%]), followed by paracorporeal pulsatile (342 [28%]), and paracorporeal continuous (327 [27%]). Overall freedom from stroke at 6 months was higher in the recent era (2012-2016; 80.2% [95% CI, 77.1%-82.9%] vs 2017-2023; 87.9% [95% CI, 86.2%-89.4%], P = .009). Implantable continuous VADs had the highest freedom from stroke at 3 months (92.7%; 95% CI, 91.1%-93.9%) and 6 months (91.1%; 95% CI, 89.3%-92.6%), followed by paracorporeal pulsatile (87.0% [95% CI, 84.8%-88.9%] and 82.8% [95% CI, 79.8%-85.5%], respectively), and paracorporeal continuous (76.0% [95% CI, 71.8%-79.5%] and 69.5% [95% CI, 63.4%-74.8%], respectively) VADs. Parametric modeling identified risk factors for stoke early after implant and later. Overall, and particularly for paracorporeal pulsatile devices, early stroke risk has decreased in the most recent era (hazard ratio, 5.01). Among implantable continuous devices, cardiogenic shock was the major risk factor. For patients <10 kg, early hazard was only seen in the previous era. For congenital patients, early hazard was seen in nonimplantable device use and use of extracorporeal membrane oxygenation., Conclusions: The overall stroke rate has decreased from 20% to 15% at 6 months, with particular improvement among paracorporeal pulsatile devices. Risk factor analyses offer insights for identification of higher stroke risk subsets and further management refinements., Competing Interests: Disclosures Joseph Rossano reports a relationship with Bayer Corporation that includes: consulting or advisory; with Merck & Co Inc that includes: consulting or advisory; with Cytokinetics Inc that includes: consulting or advisory; with Bristol-Myers Squibb Co that includes: consulting or advisory; and with American Regent Inc that includes: consulting or advisory. James Kirklin reports a relationship with The Society of Thoracic Surgeons that includes: employment. David Morales reports a relationship with Abbott that includes: consulting or advisory; with Azyio that includes: consulting or advisor; with Xeltis BV that includes: consulting or advisory; with Berlin Heart GmbH that includes: consulting or advisory; with CorMatrix Cardiovascular Inc that includes: consulting or advisory; and with SynCardia Systems LLC that includes: consulting or advisory. The other authors have no conflicts of interest to disclose., (Copyright © 2024 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Stroke affects surgical decision making and outcomes of neonatal cardiac surgery (CHS). We sought to assess the burden of stroke in this population from a large multicenter database., Methods: We analyzed neonates undergoing CHS with cardiopulmonary bypass from the Pediatric Health Information System database (2004-2022). The cohort was divided into the stroke group, which included preoperative/postoperative ischemic, hemorrhagic subtypes, and grade III to IV intraventricular hemorrhages, and compared in-hospital and follow-up outcomes to a nonstroke group., Results: A perioperative stroke occurred in 800 of 14,228 neonates (5.6%). The stroke group was more likely to have hypoplastic left heart syndrome (HLHS; 30.5% vs 20.7%), born preterm (19.4% vs 11.7%), low birth weight (17.8% vs 11.9%), and require extracorporeal membrane oxygenation (ECMO; 48.8% vs 13.8%; all P < .001). Outcomes comparing stroke vs no stroke were mortality, 33.1% vs 8.9%; nonhome discharge, 12.5% vs 6.9%; length of stay, 41 vs 24 days; and hospitalization costs, $354,521 vs $180,489 (all, P < .05). Stroke increased the odds of mortality by 2-fold (odds ratio, 2.20; 95% CI, 1.75-2.77; P < .001) after adjusting for ECMO, prematurity among other significant factors. On follow-up, the stroke group had a higher incidence of hydrocephalus (9.5% vs 1.3%), cerebral palsy (6.2% vs 1.3%), and autism spectrum disorder (7.1% vs 3.5%), and survivors of the index admission had higher 1- and 5-year mortality (5.3% and 11.3% vs 3.3% and 5.9%, respectively; all P < .05)., Conclusions: Neonatal CHS patients born prematurely, diagnosed with HLHS, or those requiring ECMO are disproportionately affected by stroke. The occurrence of stroke is marked by significantly higher mortality. Future research should seek to identify factors leading to stroke to increase rescue after stroke and for improvement of long-term outcomes., Competing Interests: Disclosures Awais Asfaq reports a relationship with PyrAmes as a primary investigator. David Morales reports consulting or advisory relationships with Abbott Vascular, Aziyo, Xeltis BV, Berlin Heart, and SynCardia Systems. The other authors have no conflicts of interest to disclose., (Copyright © 2024 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Although coarctation of the aorta without concomitant intracardiac pathology is relatively common, there is lack of guidance regarding aspects of its management in neonates and infants., Methods: A panel of experienced congenital cardiac surgeons, cardiologists, and intensivists was created, and key questions related to the management of isolated coarctation in neonates and infants were formed using the PICO (Patients/Population, Intervention, Comparison/Control, Outcome) Framework. A literature search was then performed for each question. Practice guidelines were developed with classification of recommendation and level of evidence using a modified Delphi method., Results: For neonates and infants with isolated coarctation, surgery is indicated in the absence of obvious surgical contraindications. For patients with risk factors for surgery, medical management before intervention is reasonable. For those stable off prostaglandin E 1 , the threshold for intervention remains unclear. Thoracotomy is indicated when arch hypoplasia is not present. Sternotomy is preferable when arch hypoplasia is present that cannot be adequately addressed through a thoracotomy. Sternotomy may also be considered in the presence of a bovine aortic arch. Antegrade cerebral perfusion may be reasonable when the repair is performed through a sternotomy. Extended end-to-end, arch advancement, and patch augmentation are all reasonable techniques., Conclusions: Surgery remains the standard of care for the management of isolated coarctation in neonates and infants. Depending on degree and location, arch hypoplasia may require a sternotomy approach as opposed to a thoracotomy approach. Significant opportunities remain to better delineate management in these patients., Competing Interests: Disclosures The authors have no conflicts of interest to disclose., (Copyright © 2024 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Children who undergo cardiac surgery may require post-cardiotomy extracorporeal membrane oxygenation (ECMO). Although morbidities are considerable, our understanding of outcome determinants is limited. We evaluated associations between patient and peri-operative factors with outcomes., Methods: The STS Congenital Heart Surgery Database was queried for patients <18yo who underwent post-cardiotomy ECMO from 1/2016-6/2021., Primary Outcome: survival to hospital discharge., Secondary Outcome: survival without neurologic injury. Logistic regression for binary outcomes and competing risk analysis for survival were used to identify the most important predictors. Variables were selected by stepwise procedure using entry level p=0.35. Those with p≤0.1 were kept in the final model., Results: A total of 3,181 patients were supported with post-cardiotomy ECMO during the same hospitalization as cardiac surgery: (A) intra-operative initiation of ECMO, n=1206; (B) early post-operative (≤48 hours), n=936; (C) late post-operative (>48 hours), n=1039. Most common primary procedure of index operation was the Norwood procedure. 57% intra-op survived to discharge, versus 59% early post-op and 42% late post-op group (χ 2 (2) = 64, p<0.0001, V = 0.14). In all groups, post-op septicemia, cardiac arrest, and new neurologic injury had the strongest association with mortality, while post-operative reintubation and unplanned non-cardiac reoperation were associated with higher survival., Conclusions: Multiple risk factors impact survival in children who undergo cardiac surgery and post-cardiotomy ECMO. ECMO initiated >48 hours after surgery is associated with the poorest outcomes. This is the first step in creating a predictive tool to educate clinicians and families regarding expectations in this high-risk population., (Copyright © 2024. Published by Elsevier Inc.)

Background: Recently, several centers in the United States have begun performing donation after circulatory death (DCD) heart transplants (HTs) in adults. We sought to characterize the recent use of DCD HT, waitlist time, and outcomes compared to donation after brain death (DBD)., Methods: Using the United Network for Organ Sharing database, 10,402 adult (aged >18 years) HT recipients from January 2019 to June 2022 were identified: 425 (4%) were DCD and 9,977 (96%) were DBD recipients. Posttransplant outcomes in matched and unmatched cohorts and waitlist times were compared between groups., Results: DCD and DBD recipients had similar age (57 years for both, p = 0.791). DCD recipients were more likely White (67% vs 60%, p = 0.002), on left ventricular assist device (LVAD; 40% vs 32%, p < 0.001), and listed as status 4 to 6 (60% vs 24%, p < 0.001); however, less likely to require inotropes (22% vs 40%, p < 0.001) and preoperative extracorporeal membrane oxygenation (0.9% vs 6%, p < 0.001). DCD donors were younger (29 vs 32 years, p < 0.001) and had less renal dysfunction (15% vs 39%, p < 0.001), diabetes (1.9% vs 3.8%, p = 0.050), or hypertension (9.9% vs 16%, p = 0.001). In matched and unmatched cohorts, early survival was similar (p = 0.22). Adjusted waitlist time was shorter in DCD group (21 vs 31 days, p < 0.001) compared to DBD cohort and 5-fold shorter (DCD: 22 days vs DBD: 115 days, p < 0.001) for candidates in status 4 to 6, which was 60% of DCD cohort., Conclusions: The community is using DCD mostly for those recipients who are expected to have extended waitlist times (e.g., durable LVADs, status >4). DCD recipients had similar posttransplant early survival and shorter adjusted waitlist time compared to DBD group. Given this early success, efforts should be made to expand the donor pool using DCD, especially for traditionally disadvantaged recipients on the waitlist., (Copyright © 2023 International Society for the Heart and Lung Transplantation. Published by Elsevier Inc. All rights reserved.)

In the US, the first pediatric donation after circulatory death (DCD) thoracic transplant was done in 2004; however, ethical controversy led to minimal utilization of these donors. The present study was performed to characterize the current state of pediatric DCD heart and lung transplantation (HTx, LTx). Children (<18 year old) who underwent HTx or LTx using DCD donors from June 2004 to June 2022 were identified in the United Network for Organ Sharing registry. A total of 14 DCD recipients were identified: 7 (50%) HTx and 7 (50%) LTx. Donor and recipient demographics are described in Table 1. One and 5-year post-transplant survival were as follows: HTx recipients (64% for each) and LTx recipients (86%, 55%). Although often discussed, the national experience with DCD donors for pediatric HTx and LTx remains limited and not being practiced consistently by any pediatric program. Given the critical organ shortage, DCD use in the field of pediatric thoracic transplantation should be strongly considered., (Copyright © 2023 International Society for the Heart and Lung Transplantation. Published by Elsevier Inc. All rights reserved.)

Background: The Pediatric Interagency Registry for Mechanical Circulatory Support (Pedimacs) provides detailed understanding on pediatric patients supported with ventricular assist devices (VADs). We sought to identify important variables affecting mortality in pediatric VADs., Methods: Patients aged <19 years, from 2012 to 2021, were included. Survival analyses were performed using Kaplan-Meier. Parametric hazard modeling was used to identify risk factors for death., Results: Of the 1109 patients, the most common devices were implantable continuous (IC, 448 [40%]), followed by paracorporeal pulsatile (PP, 306 [28%]), paracorporeal continuous (PC, 293 [26%]), and percutaneous (58 [5%]). Patients with percutaneous device, infants, congenital heart disease, biventricular support, and Interagency Registry for Mechanically Assisted Circulatory Support profile 1 had worse overall survival at 6 months. Positive outcome was 83% at 6 months. Consistent with their cohort composition, device type positive outcomes at 6 months were IC, 92%; PP, 84%; and PC, 69%. Parametric hazard modeling for overall survival showed an early hazard for death with biventricular support, congenital heart disease (CHD), intubation before implantation, PC device, and renal impairment, whereas a constant hazard was associated with ascites. For patients <10 kg, parametric modeling showed an early hazard for CHD, intubation, and renal impairment. Modeling in CHD patients showed an early hazard for biventricular support, renal impairment, and use of PC/PP devices., Conclusions: This multivariable analysis of the complete Pedimacs database demonstrates that illness at VAD implantation, diagnosis, and strategy of support affect survival and differ by device type. We hope this is the first step in creating a predictive tool to help providers and families have informed expectations., (Copyright © 2023 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Impact of pretransplantation risk factors on mortality in the first year after heart transplantation remains largely unknown. Using machine learning algorithms, we selected clinically relevant identifiers that could predict 1-year mortality after pediatric heart transplantation., Methods: Data were obtained from the United Network for Organ Sharing Database for years 2010-2020 for patients 0-17 years receiving their first heart transplant (N = 4150). Features were selected using subject experts and literature review. Scikit-Learn, Scikit-Survival, and Tensorflow were used. A train:test split of 70:30 was used. N-repeated k-fold validation was performed (N = 5, k = 5). Seven models were tested, Hyperparameter tuning performed using Bayesian optimization and the concordance index (C-index) was used for model assessment., Results: A C-index above 0.6 for test data was considered acceptable for survival analysis models. C-indices obtained were 0.60 (Cox proportional hazards), 0.61 (Cox with elastic net), 0.64 (gradient boosting), 0.64 (support vector machine), 0.68 (random forest), 0.66 (component gradient boosting), and 0.54 (survival trees). Machine learning models show an improvement over the traditional Cox proportional hazards model, with random forest performing the best on the test set. Analysis of the feature importance for the gradient boosted model found that the top 5 features were the most recent serum total bilirubin, the travel distance from the transplant center, the patient body mass index, the deceased donor terminal Serum glutamic pyruvic transaminase/Alanine transaminase (SGPT/ALT), and the donor PCO 2 ., Conclusions: Combination of machine learning and expert-based methodology of selecting predictors of survival for pediatric heart transplantation provides a reasonable prediction of 1- and 3-year survival outcomes. SHapley Additive exPlanations can be an effective tool for modeling and visualizing nonlinear interactions., (Copyright © 2023 International Society for the Heart and Lung Transplantation. Published by Elsevier Inc. All rights reserved.)

Background: Machine learning (ML) is an emerging tool for predicting need of end-of-life discussion and palliative care, by using mortality as a proxy. But deaths, unforeseen by emergency physicians at time of the emergency department (ED) visit, might have a weaker association with the ED visit., Objectives: To develop an ML algorithm that predicts unsurprising deaths within 30 days after ED discharge., Methods: In this retrospective registry study, we included all ED attendances within the Swedish region of Halland in 2015 and 2016. All registered deaths within 30 days after ED discharge were classified as either "surprising" or "unsurprising" by an adjudicating committee with three senior specialists in emergency medicine. ML algorithms were developed for the death subclasses by using Logistic Regression (LR), Random Forest (RF), and Support Vector Machine (SVM)., Results: Of all 30-day deaths (n = 148), 76% (n = 113) were not surprising to the adjudicating committee. The most common diseases were advanced stage cancer, multidisease/frailty, and dementia. By using LR, RF, and SVM, mean area under the receiver operating characteristic curve (ROC-AUC) of unsurprising deaths in the test set were 0.950 (SD 0.008), 0.944 (SD 0.007), and 0.949 (SD 0.007), respectively. For all mortality, the ROC-AUCs for LR, RF, and SVM were 0.924 (SD 0.012), 0.922 (SD 0.009), and 0.931 (SD 0.008). The difference in prediction performance between all and unsurprising death was statistically significant (P < .001) for all three models., Conclusion: In patients discharged to home from the ED, three-quarters of all 30-day deaths did not surprise an adjudicating committee with emergency medicine specialists. When only unsurprising deaths were included, ML mortality prediction improved significantly., (Copyright © 2021 The Author(s). Published by Elsevier Inc. All rights reserved.)

Unscheduled 30-day readmissions are a hallmark of Congestive Heart Failure (CHF) patients that pose significant health risks and escalate care cost. In order to reduce readmissions and curb the cost of care, it is important to initiate targeted intervention programs for patients at risk of readmission. This requires identifying high-risk patients at the time of discharge from hospital. Here, using real data from over 7500 CHF patients hospitalized between 2012 and 2016 in Sweden, we built and tested a deep learning framework to predict 30-day unscheduled readmission. We present a cost-sensitive formulation of Long Short-Term Memory (LSTM) neural network using expert features and contextual embedding of clinical concepts. This study targets key elements of an Electronic Health Record (EHR) driven prediction model in a single framework: using both expert and machine derived features, incorporating sequential patterns and addressing the class imbalance problem. We evaluate the contribution of each element towards prediction performance (ROC-AUC, F1-measure) and cost-savings. We show that the model with all key elements achieves higher discrimination ability (AUC: 0.77; F1: 0.51; Cost: 22% of maximum possible savings) outperforming the reduced models in at least two evaluation metrics. Additionally, we present a simple financial analysis to estimate annual savings if targeted interventions are offered to high risk patients., (Copyright © 2019 The Authors. Published by Elsevier Inc. All rights reserved.)

Background: Minimally invasive repair of pectus excavatum (MIRPE) has been extended to repair of defects in adults, with reported higher complication rates and outcome failures. The optimal revision for a prior failed MIRPE in an adult has not been identified. We review our experience for this adult population., Methods: A retrospective review was performed of 129 patients (age ≥18 years) who underwent revision after a failed pectus excavatum (PE) repair from December 2010 through December 2016., Results: In total, 47 of the 129 (36%) revision patients had a prior failed MIRPE, with 98% presenting for revision because of inadequate correction after their initial repair. The median age was 28 years (range, 18 to 54 years), and 77% were men. Thirty-one (66%) patients had indwelling pectus support bars at the revision procedure. Mean time from initial MIRPE to the revision procedure was 3.34 ± 2.9 years. A modified MIRPE was successful in 39 (83%) patients. Hybrid repair with the addition of osteotomy cuts and/or titanium plating was required in 8 patients for an adequate revision. Multiple (2 bars, 62%; 3 bars, 38%), shorter (median, 13.5 inches versus 15.0 inches) bars were used for the revision versus earlier repair. Substantial lysis of intrathoracic adhesions was required in 40 (85%) patients, with a median operative time for revision MIRPE of 169 ± 66 minutes; median operative time for hybrid procedures, 314 ± 74 minutes., Conclusions: A modified MIRPE can be successfully used in most adults to revise a failed prior MIRPE., (Copyright © 2018 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)