A Deep Learning Framework for Predicting Patient Decannulation on Extracorporeal Membrane Oxygenation Devices: Development and Model Analysis Study.

BACKGROUND: Venovenous extracorporeal membrane oxygenation (VV-ECMO) is a therapy for patients with refractory respiratory failure. The decision to decannulate someone from extracorporeal membrane oxygenation (ECMO) often involves weaning trials and clinical intuition. To date, there are limited prognostication metrics to guide clinical decision-making to determine which patients will be successfully weaned and decannulated. OBJECTIVE: This study aims to assist clinicians with the decision to decannulate a patient from ECMO, using Continuous Evaluation of VV-ECMO Outcomes (CEVVO), a deep learning-based model for predicting success of decannulation in patients supported on VV-ECMO. The running metric may be applied daily to categorize patients into high-risk and low-risk groups. Using these data, providers may consider initiating a weaning trial based on their expertise and CEVVO. METHODS: Data were collected from 118 patients supported with VV-ECMO at the Columbia University Irving Medical Center. Using a long short-term memory-based network, CEVVO is the first model capable of integrating discrete clinical information with continuous data collected from an ECMO device. A total of 12 sets of 5-fold cross validations were conducted to assess the performance, which was measured using the area under the receiver operating characteristic curve (AUROC) and average precision (AP). To translate the predicted values into a clinically useful metric, the model results were calibrated and stratified into risk groups, ranging from 0 (high risk) to 3 (low risk). To further investigate the performance edge of CEVVO, 2 synthetic data sets were generated using Gaussian process regression. The first data set preserved the long-term dependency of the patient data set, whereas the second did not. RESULTS: CEVVO demonstrated consistently superior classification performance compared with contemporary models (P<.001 and P=.04 compared with the next highest AUROC and AP). Although the model's patient-by-patient predictive power may be too low to be integrated into a clinical setting (AUROC 95% CI 0.6822-0.7055; AP 95% CI 0.8515-0.8682), the patient risk classification system displayed greater potential. When measured at 72 hours, the high-risk group had a successful decannulation rate of 58% (7/12), whereas the low-risk group had a successful decannulation rate of 92% (11/12; P=.04). When measured at 96 hours, the high- and low-risk groups had a successful decannulation rate of 54% (6/11) and 100% (9/9), respectively (P=.01). We hypothesized that the improved performance of CEVVO was owing to its ability to efficiently capture transient temporal patterns. Indeed, CEVVO exhibited improved performance on synthetic data with inherent temporal dependencies (P<.001) compared with logistic regression and a dense neural network. CONCLUSIONS: The ability to interpret and integrate large data sets is paramount for creating accurate models capable of assisting clinicians in risk stratifying patients supported on VV-ECMO. Our framework may guide future incorporation of CEVVO into more comprehensive intensive care monitoring systems.

How I manage differential gas exchange in peripheral venoarterial extracorporeal membrane oxygenation.

Dual circulation is a common but underrecognized physiological occurrence associated with peripheral venoarterial extracorporeal membrane oxygenation (ECMO). Competitive flow will develop between blood ejected from the heart and blood travelling retrograde within the aorta from the ECMO reinfusion cannula. The intersection of these two competitive flows is referred to as the "mixing point". The location of this mixing point, which depends upon the relative strengths of the native and extracorporeal pumps, will determine which regions of the body are perfused with blood ejected from the left ventricle and which regions are perfused by reinfused blood from the ECMO circuit, effectively establishing dual circulations. Because gas exchange within these circulations is dictated by the native lungs and membrane lung, respectively, oxygenation and carbon dioxide removal may differ between regions-depending on how well gas exchange is preserved within each circulation-potentially leading to differential oxygenation or differential carbon dioxide, each of which may have important clinical implications. In this perspective, we address the identification and management of dual circulation and differential gas exchange through various clinical scenarios of venoarterial ECMO. Recognition of dual circulation, proper monitoring for differential gas exchange, and understanding the various strategies to resolve differential oxygenation and carbon dioxide may allow for more optimal patient management and improved clinical outcomes.

The Use of Extracorporeal Membrane Oxygenation for COVID-19: Lessons Learned.

The coronavirus disease 2019 (COVID-19) pandemic has seen an increase in global cases of severe acute respiratory distress syndrome (ARDS), with a concomitant increased demand for extracorporeal membrane oxygenation (ECMO). Outcomes of patients with severe ARDS due to COVID-19 infection receiving ECMO support are evolving. The need for surge capacity, practical and ethical limitations on implementing ECMO, and the prolonged duration of ECMO support in patients with COVID-19-related ARDS has revealed limitations in organization and resource utilization. Coordination of efforts at multiple levels, from research to implementation, resulted in numerous innovations in the delivery of ECMO.

Optimizing Perioperative Renal Replacement Therapy for Patients Undergoing Cardiac Surgical Procedures Requiring Cardiopulmonary Bypass.

Coronary artery disease is highly prevalent in patients with end-stage renal disease (ESRD), and cardiovascular complications remain the most common cause of death in this patient population. Accordingly, many cardiac surgical procedures requiring cardiopulmonary bypass support are performed on these patients each year, with morbidity and mortality rates far exceeding patients without ESRD. Anuric patients lack the normal renal homeostatic functions which typically allow for physiologic protection from challenges during the operation, such as volume overload, hyperkalemia, and acidemia. Careful preoperative planning and coordination to provide pre-, intra-, and postoperative renal replacement therapies for such patients are imperative. Many different strategies have been reported in the literature. Zero-balance ultrafiltration is a newer strategy which utilizes convective ultrafiltration much like pre-filter continuous renal replacement therapy and utilizes pre-existing connections on the cardiopulmonary bypass pump performed by the perfusion team. This allows for control of potassium concentration throughout the operation with existing personnel and minimal additional equipment. Here, we describe the unique challenges caring for patients receiving renal replacement therapy undergoing cardiac surgical procedures requiring cardiopulmonary bypass.

Extracorporeal Membrane Oxygenation for Coronavirus Disease 2019: Crisis Standards of Care.

The coronavirus disease 2019 (COVID-19) pandemic has placed extraordinary strain on global healthcare systems. Use of extracorporeal membrane oxygenation (ECMO) for patients with severe respiratory or cardiac failure attributed to COVID-19 has been debated due to uncertain survival benefit and the resources required to safely deliver ECMO support. We retrospectively investigated adult patients supported with ECMO for COVID-19 at our institution during the first 80 days following New York City's declaration of a state of emergency. The primary objective was to evaluate survival outcomes in patients supported with ECMO for COVID-19 and describe the programmatic adaptations made in response to pandemic-related crisis conditions. Twenty-two patients with COVID-19 were placed on ECMO during the study period. Median age was 52 years and 18 (81.8%) were male. Twenty-one patients (95.4%) had severe ARDS and seven (31.8%) had cardiac failure. Fifteen patients (68.1%) were managed with venovenous ECMO while 7 (31.8%) required arterial support. Twelve patients (54.5%) were transported on ECMO from external institutions. Twelve patients were discharged alive from the hospital (54.5%). Extracorporeal membrane oxygenation was used successfully in patients with respiratory and cardiac failure due to COVID-19. The continued use of ECMO, including ECMO transport, during crisis conditions was possible even at the height of the COVID-19 pandemic.

Discussion: can upper extremity (deltoid) near infrared spectroscopy be used to assess cerebral tissue bed saturation on femorally cannulated veno-arterial extracorporeal membrane oxygenation patients?

Continuous cerebral tissue saturation monitoring with near infrared spectroscopy may help clinicians identify cerebral desaturation early; however, patients have reported discomfort from near infrared spectroscopy monitoring pads on the forehead. This study aims to compare upper extremity near infrared spectroscopy monitoring to cerebral near infrared spectroscopy monitoring to assess its viability as a surrogate for cerebral saturation. A retrospective analysis of 10 femorally cannulated veno-arterial extracorporeal membrane oxygenation patients was performed comparing left (L) and right (R) upper extremity (deltoid) near infrared spectroscopy monitoring to cerebral near infrared spectroscopy monitoring (n = 20 data sets, 10 left and 10 right) and right radial blood gasses. Deltoid and cerebral near infrared spectroscopy values were recorded every 15 minutes for at least 24 hours when possible, were plotted on scatter grams, and were analyzed using Pearson product-moment coefficient (r). Based on the concept of covariance, a moderate-good relationship r = 0.50-0.75 was noted in 10% (n = 2) of the study group. A fair relationship r = 0.25-0.50 was noted in 50% (n = 10), and little or no relationship was noted in 40% (n = 8). None of the study group displayed a good to excellent relationship (r = 0.75 or above). In addition, coefficient of multiple determination for multiple regression R2 was calculated and strong fit of the regression line was not noted. Although cerebral near infrared spectroscopy monitoring has been extremely helpful in identifying low cerebral tissue saturation on veno-arterial extracorporeal membrane oxygenation patients, the use of upper extremity (peripheral deltoid) tissue monitoring does not provide adequate correlation and should not be used as a surrogate to cerebral monitoring.