Machine Learning from Veno-Venous Extracorporeal Membrane Oxygenation Identifies Factors Associated with Neurological Outcomes.

BACKGROUND: Neurological complications are common in patients receiving veno-venous extracorporeal membrane oxygenation (VV-ECMO) support. We used machine learning (ML) algorithms to identify predictors for neurological outcomes for these patients. METHODS: All demographic, clinical, and circuit-related variables were extracted for adults with VV-ECMO support at a tertiary care center from 2016 to 2022. The primary outcome was good neurological outcome (GNO) at discharge defined as a modified Rankin Scale of 0-3. RESULTS: Of 99 total VV-ECMO patients (median age = 48 years; 65% male), 37% had a GNO. The best performing ML model achieved an area under the receiver operating characteristic curve of 0.87. Feature importance analysis identified down-trending gas/sweep/blender flow, FiO2, and pump speed as the most salient features for predicting GNO. CONCLUSION: Utilizing pre- as well as post-initiation variables, ML identified on-ECMO physiologic and pulmonary conditions that best predicted neurological outcomes.

Changes in Donor Utilization and Outcomes for Patients Bridged With Durable Left Ventricular Assist Device.

We studied the impact of the 2018 heart allocation policy change on donor characteristics and posttransplant outcomes of left ventricular assist device (LVAD)-bridged heart transplant (HT) recipients. Left ventricular assist device-bridged adult HT recipients from October 2014 to October 2022 in the United Network for Organ Sharing database were categorized into old allocation policy (OAP) and new allocation policy (NAP) cohorts. Baseline characteristics, posttransplant outcomes, and subgroup analyses of unstable and stable LVAD-bridged recipients were assessed. The study included 7,384 HT recipients; 4,345 (58.8%) were transplanted in the OAP era and 3,039 (41.2%) in the NAP era. Old allocation policy recipients were most frequently status 1A at transplantation (71.1%), whereas NAP recipients were most frequently status 3 (40.0%), and status 4 (31.9%). Median donor sequence number (DSN) was higher in the NAP versus OAP era (9 vs. 3, p < 0.001). On multivariable analysis, NAP recipients had 20% higher 1 year mortality compared to OAP (adjusted hazard ratio [aHR] = 1.20 [95% confidence interval {CI}: 1.04-1.40], p = 0.01). Status 1 or 2 recipients had 28% higher 1 year mortality compared to status 1A (aHR = 1.28 [95% CI: 1.01-1.63], p = 0.04). Status 1 and 2 LVAD-supported recipients had higher mortality following the 2018 allocation change, indicating the need for closer surveillance of LVAD-bridged patients who may decompensate on the waitlist.

Impact of Extracorporeal Membrane Oxygenation Bridging Duration on Lung Transplant Outcomes.

BACKGROUND: We sought to characterize the association between venovenous extracorporeal membrane oxygenation (VV-ECMO) bridging duration and outcomes in patients listed for lung transplantation. METHODS: A retrospective observational study was conducted using the Organ Procurement and Transplantation Network (OPTN) database to identify adults (aged ≥18 years) who were listed for lung transplantation between 2016 and 2020 and were bridged with VV-ECMO. Patients were then stratified into groups, determined by risk inflection points, depending on the amount of time spent on pretransplant ECMO: group 1 (≤5 days), group 2 (6-10 days), group 3 (11-20 days), and group 4 (>20 days). Waiting list survival between groups was analyzed using Fine-Gray competing risk models. Posttransplant survival was compared using Cox regression. RESULTS: Of 566 eligible VV-ECMO bridge-to-lung-transplant patients (median age, 54 years, 49% men), 174 (31%), 124 (22%), 130 (23%), and 138 (24%) were categorized as groups 1, 2, 3, and 4, respectively. Overall, median duration of VV-ECMO was 10 days (interquartile range, 1-211 days), and 178 patients (31%) died on the waiting list. In the Fine-Gray model, compared with group 1, patients bridged with longer ECMO durations in group 2 (subdistribution hazard ratio [SHR], 2.95; 95% CI, 1.63-5.35), group 3 (SHR, 3.96; 95% CI, 2.36-6.63), and group 4 (SHR, 4.33; 95% CI, 2.59-7.22, all P < .001) were more likely to die on the waiting list. Of 388 patients receiving a transplant, pretransplant ECMO duration was not associated with 1-year survival in Cox regression. CONCLUSIONS: Prolonged duration of ECMO bridging was associated with worse waiting list mortality but did not impact survival after lung transplant. Prioritization of very early transplantation may improve waiting list outcomes in this population.

Bigger pies, bigger slices: Increased hospitalization costs for lung transplantation recipients in the non-donation service area allocation era.

OBJECTIVE: On November 24, 2017, lung transplant allocation switched from donation service area to a 250-nautical mile radius policy to improve equity in access to lung transplantation. Given the growing consideration of healthcare costs, we evaluated changes in hospitalization costs after this policy change. METHODS: Lung transplant hospitalizations were identified within the National Inpatient Sample from 2005 to 2020. Recipients were categorized as donation service area era (August 2015 to October 2017) or non-donation service area era (December 2017 to February 2020). Median total hospitalization costs (inflation adjusted) were compared by era nationally and regionally. Multivariable generalized linear regression was performed to determine if the removal of the donation service area was associated with total hospitalization costs. The model was adjusted for recipient demographics, Charlson Comorbidity Index, hospitalization region, transplant type (single, double), and use of extracorporeal membrane oxygenation, ex vivo lung perfusion, and mechanical ventilation. RESULTS: We analyzed 12,985 lung transplant recipients (median age of 61 years, 66% were male): 7070 in the donation service area era and 5915 in the non-donation service area era. Demographics were not different between recipients in both eras. Non-donation service area era recipients had greater extracorporeal membrane oxygenation use, mechanical ventilation (<24 hours), and longer length of stay than donation service area era recipients. Median total hospitalization costs for non-donation service area versus donation service area era recipients increased by $24,198 ($157,964 vs $182,162, percentage change = 15.32%, P < .001). Median costs increased in East North Central ($42,281) and Mountain ($35,521) regions (both P < .01). After adjustment, median costs for non-donation service area versus donation service area era recipients still increased ($19,168, 95% CI, 145-38,191, P = .048). CONCLUSIONS: Hospitalization costs for lung transplant hospitalizations have increased from 2015 to 2020. The transition from donation service area-based allocation to the non-donation service area system may have contributed to this increase after 2017 by increasing access to transplant for sicker recipients.

The Impact of Local Programmatic Decisions on Outcomes in Transplant-Listed Adults With Congenital Heart Disease.

BACKGROUND: We investigated variables impacting waitlist times and negative waitlist outcomes in adults with congenital heart disease (ACHD) who were waiting for orthotopic heart transplant (OHT) after the 2018 allocation change. METHODS: Adult candidates for OHT who were listed between 10/18/2018 and 12/31/2022 in the United Network for Organ Sharing database were categorized as ACHD vs non-ACHD. Waitlist time and time to upgrade for those upgraded into status 1-3 were compared by using rank-sum tests. Death/delisting for deterioration was assessed by using Fine-Gray subdistribution hazard ratios (SHRs). RESULTS: Of 15,424 OHT candidates, 589 (3.8%) were ACHD. ACHD vs non-ACHD candidates had less urgent status at initial listing (4.2% vs 4.7% listed at status 1; 17.2% vs 23.7% listed at status 2; P < 0.001), but not final listing (5.9% vs 7.6% final status 1; 35.6% vs 36.8% final status 2; P < 0.001). ACHD vs non-ACHD candidates upgraded into status 1 (65.0 vs 30.0 days; P = 0.09) and status 2 (113.0 vs 64.0 days; P = 0.003) spent longer times on the waitlist. ACHD vs non-ACHD candidates spent longer times waiting for an upgrade into status 1 (51.4 vs 17.6 days; P = 0.027) and status 2 (76.7 vs 34.7 days; P = 0.003). Once upgraded, there was no difference between groups in waitlist time to status 1 (9.7 vs 5.5 days = 0.66). ACHD vs non-ACHD candidates with a final status of 1 (20.0% vs 8.6%; SHR 2.47 [95%CI = 1.19-5.16]; P = 0.02) and 2 (8.9% vs 2.3%; SHR 3.59 [95%CI = 2.18-5.91]; P < 0.001) experienced higher rates of death and deterioration. CONCLUSIONS: ACHD candidates have longer waitlist times, have lower priority status at initial listing, wait longer for upgrades, and have higher mortality rates at the same final status as non-ACHD candidates, suggesting that they are being upgraded too late.

Predicting Acute Brain Injury in Venoarterial Extracorporeal Membrane Oxygenation Patients with Tree-Based Machine Learning: Analysis of the Extracorporeal Life Support Organization Registry.

Objective: To determine if machine learning (ML) can predict acute brain injury (ABI) and identify modifiable risk factors for ABI in venoarterial extracorporeal membrane oxygenation (VA-ECMO) patients. Design: Retrospective cohort study of the Extracorporeal Life Support Organization (ELSO) Registry (2009-2021). Setting: International, multicenter registry study of 676 ECMO centers. Patients: Adults (≥18 years) supported with VA-ECMO or extracorporeal cardiopulmonary resuscitation (ECPR). Interventions: None. Measurements and Main Results: Our primary outcome was ABI: central nervous system (CNS) ischemia, intracranial hemorrhage (ICH), brain death, and seizures. We utilized Random Forest, CatBoost, LightGBM and XGBoost ML algorithms (10-fold leave-one-out cross-validation) to predict and identify features most important for ABI. We extracted 65 total features: demographics, pre-ECMO/on-ECMO laboratory values, and pre-ECMO/on-ECMO settings.Of 35,855 VA-ECMO (non-ECPR) patients (median age=57.8 years, 66% male), 7.7% (n=2,769) experienced ABI. In VA-ECMO (non-ECPR), the area under the receiver-operator characteristics curves (AUC-ROC) to predict ABI, CNS ischemia, and ICH was 0.67, 0.67, and 0.62, respectively. The true positive, true negative, false positive, false negative, positive, and negative predictive values were 33%, 88%, 12%, 67%, 18%, and 94%, respectively for ABI. Longer ECMO duration, higher 24h ECMO pump flow, and higher on-ECMO PaO2 were associated with ABI.Of 10,775 ECPR patients (median age=57.1 years, 68% male), 16.5% (n=1,787) experienced ABI. The AUC-ROC for ABI, CNS ischemia, and ICH was 0.72, 0.73, and 0.69, respectively. The true positive, true negative, false positive, false negative, positive, and negative predictive values were 61%, 70%, 30%, 39%, 29% and 90%, respectively, for ABI. Longer ECMO duration, younger age, and higher 24h ECMO pump flow were associated with ABI. Conclusions: This is the largest study predicting neurological complications on sufficiently powered international ECMO cohorts. Longer ECMO duration and higher 24h pump flow were associated with ABI in both non-ECPR and ECPR VA-ECMO.

Outcomes of Recipients Aged 65 Years and Older Bridged to Lung Transplant With Extracorporeal Membrane Oxygenation.

Extracorporeal membrane oxygenation (ECMO) as a bridge to lung transplant (BTT) has been used for critically ill candidates with excellent outcomes, but data on this strategy in older recipients remain limited. We compared outcomes of no BTT, mechanical ventilation (MV)-only BTT, and ECMO BTT in recipients of greater than or equal to 65 years. Lung-only recipients of greater than or equal to 65 years in the United Network for Organ Sharing database between 2008 and 2022 were included and stratified by bridging strategy. Of the 9,936 transplants included, 226 (2.3%) were MV-only BTT and 159 (1.6%) were ECMO BTT. Extracorporeal membrane oxygenation BTT recipients were more likely to have restrictive disease pathology, had higher median lung allocation score, and spent fewer days on the waitlist (all p < 0.001). Compared to no-BTT recipients, ECMO BTT recipients were more likely to be intubated or on ECMO at 72 hours posttransplant and had longer hospital lengths of stay (all p < 0.001). Extracorporeal membrane oxygenation BTT recipients had increased risk of 3 years mortality compared to both no-BTT (adjusted hazard ratio [aHR] = 1.48 [95% confidence interval {CI}: 1.14-1.91], p = 0.003) and MV-only recipients (aHR = 1.50 [95% CI: 1.08-2.07], p = 0.02). Overall, we found that ECMO BTT in older recipients is associated with inferior posttransplant outcomes compared to MV-only or no BTT, but over half of recipients remained alive at 3 years posttransplant.

Risk Factors for Nondiagnostic Imaging in a Real-World Deployment of Artificial Intelligence Diabetic Retinal Examinations in an Integrated Healthcare System: Maximizing Workflow Efficiency Through Predictive Dilation.

OBJECTIVE: In the pivotal clinical trial that led to Food and Drug Administration De Novo "approval" of the first fully autonomous artificial intelligence (AI) diabetic retinal disease diagnostic system, a reflexive dilation protocol was used. Using real-world deployment data before implementation of reflexive dilation, we identified factors associated with nondiagnostic results. These factors allow a novel predictive dilation workflow, where patients most likely to benefit from pharmacologic dilation are dilated a priori to maximize efficiency and patient satisfaction. METHODS: Retrospective review of patients who were assessed with autonomous AI at Johns Hopkins Medicine (8/2020 to 5/2021). We constructed a multivariable logistic regression model for nondiagnostic results to compare characteristics of patients with and without diagnostic results, using adjusted odds ratio (aOR). P < .05 was considered statistically significant. RESULTS: Of 241 patients (59% female; median age = 59), 123 (51%) had nondiagnostic results. In multivariable analysis, type 1 diabetes (T1D, aOR = 5.82, 95% confidence interval [CI]: 1.45-23.40, P = .01), smoking (aOR = 2.86, 95% CI: 1.36-5.99, P = .005), and age (every 10-year increase, aOR = 2.12, 95% CI: 1.62-2.77, P < .001) were associated with nondiagnostic results. Following feature elimination, a predictive model was created using T1D, smoking, age, race, sex, and hypertension as inputs. The model showed an area under the receiver-operator characteristics curve of 0.76 in five-fold cross-validation. CONCLUSIONS: We used factors associated with nondiagnostic results to design a novel, predictive dilation workflow, where patients most likely to benefit from pharmacologic dilation are dilated a priori. This new workflow has the potential to be more efficient than reflexive dilation, thus maximizing the number of at-risk patients receiving their diabetic retinal examinations.

Incidence, risk factors, and outcomes of postoperative stroke in combined heart-lung transplantation: A retrospective cohort study of the UNOS registry.

Stroke is a well-characterized complication of isolated heart and lung transplantation, but has not been described in combined heart-lung transplantation (HLTx). We retrospectively reviewed national U.S. data to describe the incidence, risk factors, and impact of postoperative stroke in HLTx recipients. Of 871 heart-lung recipients between 1994-2022, 35 (4.0%) experienced stroke, and the incidence increased over time, trending toward significance (p-trend = .07). After adjustment, extracorporeal membrane oxygenation (ECMO) (Adjusted odds ratio [aOR] = 2.63, 95%CI = [1.13-6.11]) and pre-transplant implantable defibrillator (aOR = 2.86, 95%CI = [1.20-6.81]) were independent risk factors for stroke. Postoperative stroke is common and is increasing in an era where organ allocation is driven by mechanical circulatory support (MCS) bridging.

Outcomes of Lung Transplant Candidates Aged ≥70 Years During the Lung Allocation Score Era.

BACKGROUND: With the increasing age of lung transplant candidates, we studied waitlist and posttransplantation outcomes of candidates ≥70 years during the Lung Allocation Score era. METHODS: Adult lung transplant candidates from 2005 to 2020 in the United Network for Organ Sharing database were included and stratified on the basis of age at listing into 18 to 59 years old, 60 to 69 years old, and ≥70 years old. Baseline characteristics, waitlist outcomes, and posttransplantation outcomes were assessed. RESULTS: A total of 37,623 candidates were included (52.3% aged 18-59 years, 40.6% aged 60-69 years, 7.1% aged ≥70 years). Candidates ≥70 years were more likely than younger candidates to receive a transplant (81.9% vs 72.7% [aged 60-69 years] vs 61.6% [aged 18-59 years]) and less likely to die or to deteriorate on the waitlist within 1 year (9.1% vs 10.1% [aged 60-69 years] vs 12.2% [aged 18-59 years]; P < .001). Donors for older recipients were more likely to be extended criteria (75.7% vs 70.1% [aged 60-69 years] vs 65.7% [aged 18-59 years]; P < .001). Recipients ≥70 years were found to have lower rates of acute rejection (6.7% vs 7.4% [aged 60-69 years] vs 9.2% [aged 18-59 years]; P < .001) and prolonged intubation (21.7% vs 27.4% [aged 60-69 years] vs 34.5% [aged 18-59 years]; P < .001). Recipients aged ≥70 years had increased 1-year (adjusted hazard ratio [aHR], 1.19 [95% CI, 1.06-1.33]; P < .001), 3-year (aHR, 1.28 [95% CI, 1.18-1.39]; P < .001), and 5-year mortality (aHR, 1.29 [95% CI, 1.21-1.38]; P < .001) compared with recipients aged 60 to 69 years. CONCLUSIONS: Candidates ≥70 years had favorable waitlist and perioperative outcomes despite increased use of extended criteria donors. Careful selection of candidates and postoperative surveillance may improve posttransplantation survival in this population.

National utilization, trends, and lung transplant outcomes of static versus portable ex vivo lung perfusion platforms.

BACKGROUND: This study compared utilization and outcomes of the 2 widely utilized ex vivo lung perfusion (EVLP) platforms in the United States: a static platform and a portable platform. METHODS: Adult (age 18 years or older) bilateral lung-only transplants utilizing EVLP between February 28, 2018, and December 31, 2022, in the United Network for Organ Sharing database were included. Predischarge acute rejection, intubation at 72 hours posttransplant, extracorporeal membrane oxygenation at 72 hours posttransplant, primary graft dysfunction grade 3 at 72 hours posttransplant, 30-day mortality, and 1-year mortality were evaluated using multivariable regressions. RESULTS: Overall, 607 (6.3%) lung transplants during the study period used EVLP (51.2% static, 48.8% portable). Static EVLP was primarily utilized in the eastern United States, whereas portable EVLP was primarily utilized in the western United States. Static EVLP donors were more likely to be donation after circulatory death (33.4% vs 26.0%; P = .005), have a >20 pack-year smoking history (13.5% vs 6.5%; P = .005), and be extended criteria donors (92.3% vs 85.0%; P = .013), whereas portable EVLP donors were more likely to be older than age 55 years (14.2% vs 8.0%; P = .02). Transplants utilizing the static and portable platforms had similar risk of acute rejection, intubation at 72 hours, extracorporeal membrane oxygenation at 72 hours, primary graft dysfunction grade 3 at 72 hours, and posttransplant mortality at 30 days and 1 year (all P values > .05). CONCLUSIONS: The static and portable platforms had significant differences in donor characteristics and geographic distributions of utilization. Despite this, posttransplant survival was similar between the 2 EVLP platforms.

Epidemiology of ischemic stroke and hemorrhagic stroke in venoarterial extracorporeal membrane oxygenation.

BACKGROUND: While venoarterial extracorporeal membrane oxygenation (V-A ECMO) provides lifesaving support for cardiopulmonary failure, complications may increase mortality, with few studies focusing on ischemic/hemorrhagic stroke. We aimed to determine the trends and associations of stroke incidence and mortality, and their risk factors, including the effects of annual case volumes of ECMO centers. METHODS: Retrospective analysis was performed on the Extracorporeal Life Support Organization (ELSO) registry, including adult V-A ECMO patients from 534 international centers between 2012 and 2021, excluding extracorporeal cardiopulmonary resuscitation. Temporal trend analyses were performed for stroke incidence and mortality. Univariate testing, multivariable regression, and survival analysis were used to evaluate the associations of stroke, 90-day mortality, and impact of annual center volume. RESULTS: Of 33,041 patients, 20,297 had mortality data, and 12,327 were included in the logistic regression. Between 2012 and 2021, ischemic stroke incidence increased (p < 0.0001), hemorrhagic stroke incidence remained stable, and overall 90-day mortality declined (p < 0.0001). Higher 24-h PaO2 and greater decrease between pre-ECMO PaCO2 and post-cannulation 24-h PaCO2 were associated with greater ischemic stroke incidence, while annual case volume was not. Ischemic/hemorrhagic strokes were associated with increased 90-day mortality (both p < 0.0001), while higher annual case volume was associated with lower 90-day mortality (p = 0.001). Hazard of death was highest in the first several days of V-A ECMO. CONCLUSION: In V-A ECMO patients between 2012 and 2021, 90-day mortality decreased, while ischemic stroke incidence increased. ELSO centers with higher annual case volumes had lower mortality, but were not associated with ischemic/hemorrhagic stroke incidence. Both ischemic/hemorrhagic strokes were associated with increased mortality.

Impact of intraoperative blood products, fluid administration, and persistent hypothermia on bleeding leading to reexploration after cardiac surgery.

OBJECTIVE: Risk factors for severe postoperative bleeding after cardiac surgery remain multiple and incompletely elucidated. We evaluated the impact of intraoperative blood product transfusions, intravenous fluid administration, and persistently low core body temperature (CBT) at intensive care unit arrival on risk of perioperative bleeding leading to reexploration. METHODS: We retrospectively queried our tertiary care center's Society of Thoracic Surgeons Institutional Database for all index, on-pump, adult cardiac surgery patients between July 2016 and September 2022. Intraoperative fluid (crystalloid and colloid) and blood product administrations, as well as perioperative CBT data, were harvested from electronic medical records. Linear and nonlinear mixed models, treating surgeon as a random effect to account for inter-surgeon practice differences, were used to assess the association between above factors and reexploration for bleeding. RESULTS: Of 4037 patients, 151 (3.7%) underwent reexploration for bleeding. Reexplored patients experienced remarkably greater postoperative morbidity (23% vs 6%, P < .001) and 30-day mortality (14% vs 2%, P < .001). In linear models, progressively increasing IV crystalloid administration (adjusted odds ratio, 1.11, 95% confidence interval, 1.03-1.19) and decreasing CBT on intensive care unit arrival (adjusted odds ratio, 1.20; 95% confidence interval, 1.05-1.37) were associated with greater risk of bleeding leading to reexploration. Nonlinear analysis revealed increasing risk after ∼6 L of crystalloid administration and a U-shaped relationship between CBT and reexploration risk. Intraoperative blood product transfusion of any kind was not associated with reexploration. CONCLUSIONS: We found evidence of both dilution- and hypothermia-related effects associated with perioperative bleeding leading to reexploration in cardiac surgery. Interventions targeting modification of such risk factors may decrease the rate this complication.

What Factors Are Associated With Arterial Line-Related Limb Ischemia in Patients on Extracorporeal Membrane Oxygenation? A Single-Center Retrospective Cohort Study.

OBJECTIVES: The primary purpose of this study was to identify factors associated with the development of arterial line-related limb ischemia in patients on extracorporeal membrane oxygenation (ECMO). The authors also sought to characterize and report the outcomes of patients who developed arterial line-related limb ischemia. DESIGN: Retrospective cohort study. SETTING: A single academic tertiary referral ECMO center. PARTICIPANTS: Consecutive patients who were treated with ECMO over 6 years. INTERVENTIONS: Use of arterial line. MEASUREMENTS AND MAIN RESULTS: A total of 278 consecutive ECMO patients were included, with 19 (7%) patients developing arterial line-related limb ischemia during the ECMO run. Postcannulation Sequential Organ Failure Assessment (SOFA) (adjusted odds ratio [aOR] 1.20, 95% CI 1.08-1.32), Acute Physiology and Chronic Health Evaluation-II (aOR 0.84, 95% CI 0.74-0.95), and adjusted Vasopressor Dose Equivalence (aOR 1.03, 95% CI 1.01-1.05) scores were independently associated with the development of arterial line-associated limb ischemia. A SOFA score of ≥17 at the time of ECMO cannulation had an 80% sensitivity and 87% specificity for predicting arterial line-related limb ischemia. CONCLUSIONS: Arterial line-related limb ischemia is much more common in ECMO patients than in the typical intensive care unit setting. The SOFA score may be useful in identifying which patients may be at risk for arterial line-related limb ischemia. As this was a single-center retrospective study, these results are inherently exploratory, and prospective multicenter studies are necessary to validate these results.

Extracorporeal Membrane Oxygenation Physiological Factors Influence Pulse Oximetry and Arterial Oxygen Saturation Discrepancies.

BACKGROUND: Cannulation strategy, vasopressors, and hemolysis are important physiological factors that influence hemodynamics in extracorporeal membrane oxygenation (ECMO). We hypothesized these factors influence the discrepancy between oxygen saturation measured by pulse oximetry (Spo2) and arterial blood gas (Sao2) in patients on ECMO. METHODS: We retrospectively analyzed adults (aged ≥18 years) on venoarterial or venovenous ECMO at a tertiary academic ECMO center. Spo2-Sao2 pairs with oxygen saturation ≥70% and measured within 10 minutes were included. Occult hypoxemia was defined as Sao2 ≤88% with a time-matched Spo2 ≥92%. Adjusted linear mixed-effects modeling was used to assess the Spo2-Sao2 discrepancy with preselected demographics and time-matched laboratory variables. Vasopressor use was quantified by vasopressor dose equivalences. RESULTS: Of 139 venoarterial-ECMO and 88 venovenous-ECMO patients, we examined 20,053 Spo2-Sao2 pairs. The Spo2-Sao2 discrepancy was greater in venovenous-ECMO (1.15%) vs venoarterial-ECMO (-0.35%, P < .001). Overall, 81 patients (35%) experienced occult hypoxemia during ECMO. Occult hypoxemia was more common in venovenous-ECMO (65%) than in venoarterial-ECMO (17%, P < .001). In linear mixed-effects modeling, Spo2 underestimated Sao2 by 9.48% in central vs peripheral venoarterial-ECMO (95% CI, -17.1% to -1.79%; P = .02). Higher vasopressor dose equivalences significantly worsened the Spo2-Sao2 discrepancy (P < .001). In linear mixed-effects modeling, Spo2 overestimated Sao2 by 25.43% in single lumen-cannulated vs double lumen-cannulated venovenous-ECMO (95% CI, 5.27%-45.6%; P = .03). Higher vasopressor dose equivalences and lactate dehydrogenase levels significantly worsened the Spo2-Sao2 discrepancy (P < .001). CONCLUSIONS: Venovenous-ECMO patients are at higher risk for occult hypoxemia compared with venoarterial-ECMO. A higher vasopressor requirement and different cannulation strategies (central venoarterial-ECMO; single-lumen venovenous-ECMO) were significant factors for clinically significant Spo2-Sao2 discrepancy in both ECMO modes.

Trends on Near-Infrared Spectroscopy Associated With Acute Brain Injury in Venoarterial Extracorporeal Membrane Oxygenation.

We aimed to determine the association between cerebral regional oxygen saturation (rSO 2 ) trends from cerebral near-infrared spectroscopy (cNIRS) and acute brain injury (ABI) in adult venoarterial extracorporeal membrane oxygenation (VA-ECMO) patients. ABI was defined as intracranial hemorrhage, ischemic stroke, hypoxic ischemic brain injury, or brain death during ECMO. rSO 2 values were collected from left and right cerebral oximetry sensors every hour from ECMO cannulation. Cerebral desaturation was defined as consecutive hours of rSO 2 < 40%. rSO 2 asymmetry was determined by (a) averaging left/right rSO 2 difference over the entire ECMO run; (b) consecutive hours of rSO 2 asymmetry. Sixty-nine VA-ECMO patients (mean age 56 years, 65% male) underwent cNIRS. Eighteen (26%) experienced ABI. When the mean rSO 2 asymmetry was >8% there was significantly increased odds of ABI (aOR = 39.4; 95% CI = 4.1-381.4). Concurrent rSO 2 < 40% and rSO 2 asymmetry >10% for >10 consecutive hours (asymmetric desaturation) was also significantly associated with ABI (aOR = 5.2; 95% CI = 1.2-22.2), but neither criterion alone were. Mean rSO 2 asymmetry>8% exhibited 39% sensitivity and 98% specificity for detecting ABI, with an area under the curve (AUC) of 0.86, and asymmetric desaturation had 33% sensitivity and 88% specificity, with an AUC of 0.72. These trends on NIRS monitoring may help detect ABI in VA-ECMO patients.

Epidemiology of Ischemic Stroke and Hemorrhagic Stroke in Venoarterial Extracorporeal Membrane Oxygenation.

Background: While venoarterial extracorporeal membrane oxygenation (VA-ECMO) provides lifesaving support for cardiopulmonary failure, complications may arise that increase mortality, with few studies focusing on ischemic/hemorrhagic stroke. We aimed to determine the trends of stroke incidence and mortality, associations with each other, and associations with total case volume at each Extracorporeal Life Support Organization (ELSO) center. Methods: Retrospective analysis of ELSO registry, including adult VA-ECMO patients from 534 international centers between 2012-2021, excluding extracorporeal cardiopulmonary resuscitation. Cochran-Armitage test and Poisson regression were used for trend analysis of stroke incidence and mortality. Kaplan-Meier curves, hazard functions, and multivariable logistic regression were used to study the impact of stroke on 90-day mortality. Results: Of 33,041 patients (median age = 58 years, female = 32%), 4% developed ischemic stroke, and 2% developed hemorrhagic stroke. Ischemic stroke incidence increased (×1.21/year, p < 0.0001), while hemorrhagic stroke incidence remained stable, and overall 90-day mortality declined (1.78%/year, p < 0.0001). Ischemic/hemorrhagic strokes were associated with increased overall 90-day mortality (OR = 3.29, 3.99 respectively, both p < 0.0001) after controlling for pre-selected covariates, including age, pre/post-cannulation lab values, ECMO duration, center volume, and on-ECMO complications. Total center volume was associated positively with ischemic/hemorrhagic stroke incidences (OR = 1.039, 1.053 per-additional-100-cases respectively, both p = 0.022), but inversely with 90-day mortality (OR = 0.909 per-additional-100-cases, p < 0.0001). Hazard of death was highest in the first several days of VA-ECMO. Conclusion: In VA-ECMO patients, while the reported ischemic stroke incidence steadily increased over time, 90-day mortality decreased. ELSO centers with higher case volumes reported greater stroke incidence, but lower mortality. Both ischemic/hemorrhagic strokes were associated with increased mortality.

Racial and ethnical discrepancy in hypoxemia detection in patients on extracorporeal membrane oxygenation.

Objective: To determine whether there is racial/ethnical discrepancy between pulse oximetry (SpO2) and oxygen saturation (SaO2) in patients receiving extracorporeal membrane oxygenation (ECMO). Methods: This was a retrospective observational study at a tertiary academic ECMO center with adults (>18 years) on venoarterial (VA) or venovenous (VV) ECMO. Datapoints were excluded if oxygen saturation ≤70% or SpO2-SaO2 pairs were not measured within 10 minutes. The primary outcome was the presence of a SpO2-SaO2 discrepancy between different races/ethnicities. Bland-Altman analyses and linear mixed-effects modeling, adjusting for prespecified covariates, were used to assess the SpO2-SaO2 discrepancy between races/ethnicities. Occult hypoxemia was defined as SaO2 <88% with a time-matched SpO2 ≥92%. Results: Of 139 patients receiving VA-ECMO and 57 patients receiving VV-ECMO, we examined 16,252 SpO2-SaO2 pairs. The SpO2-SaO2 discrepancy was greater in VV-ECMO (1.4%) versus VA-ECMO (0.15%). In VA-ECMO, SpO2 overestimated SaO2 in Asian (0.2%), Black (0.94%), and Hispanic (0.03%) patients and underestimated SaO2 in White (-0.06%) and nonspecified race (-0.80%) patients. The proportion of SpO2-SaO2 measurements considered occult hypoxemia was 70% from Black compared to 27% from White patients (P < .0001). In VV-ECMO, SpO2 overestimated SaO2 in Asian (1.0%), Black (2.9%), Hispanic (1.1%), and White (0.50%) patients and underestimated SaO2 in nonspecified race patients (-0.53%). In linear mixed-effects modeling, SpO2 overestimated SaO2 by 0.19% in Black patients (95% confidence interval, 0.045%-0.33%, P = .023). The proportion of SpO2-SaO2 measurements considered occult hypoxemia was 66% from Black compared with 16% from White patients (P < .0001). Conclusions: SpO2 overestimates SaO2 in Asian, Black, and Hispanic versus White patients, and this discrepancy was greater in VV-ECMO versus VA-ECMO, suggesting the need for physiological studies.

Characteristics of delirium and its association with sedation and in-hospital mortality in patients with COVID-19 on veno-venous extracorporeal membrane oxygenation.

Background: Veno-venous extracorporeal membrane oxygenation (VV-ECMO) has been used in patients with COVID-19 acute respiratory distress syndrome (ARDS). We aim to assess the characteristics of delirium and describe its association with sedation and in-hospital mortality. Methods: We retrospectively reviewed adult patients on VV-ECMO for severe COVID-19 ARDS in the Johns Hopkins Hospital ECMO registry in 2020-2021. Delirium was assessed by the Confusion Assessment Method for the ICU (CAM-ICU) when patients scored-3 or above on the Richmond Agitation-Sedation Scale (RASS). Primary outcomes were delirium prevalence and duration in the proportion of days on VV-ECMO. Results: Of 47 patients (median age = 51), 6 were in a persistent coma and 40 of the remaining 41 patients (98%) had ICU delirium. Delirium in the survivors (n = 21) and non-survivors (n = 26) was first detected at a similar time point (VV-ECMO day 9.5(5,14) vs. 8.5(5,21), p = 0.56) with similar total delirium days on VV-ECMO (9.5[3.3, 16.8] vs. 9.0[4.3, 28.3] days, p = 0.43). Non-survivors had numerically lower RASS scores on VV-ECMO days (-3.72[-4.42, -2.96] vs. -3.10[-3.91, -2.21], p = 0.06) and significantly prolonged delirium-unassessable days on VV-ECMO with a RASS of -4/-5 (23.0[16.3, 38.3] vs. 17.0(6,23), p = 0.03), and total VV-ECMO days (44.5[20.5, 74.3] vs. 27.0[21, 38], p = 0.04). The proportion of delirium-present days correlated with RASS (r = 0.64, p < 0.001), the proportions of days on VV-ECMO with a neuromuscular blocker (r = -0.59, p = 0.001), and with delirium-unassessable exams (r = -0.69, p < 0.001) but not with overall ECMO duration (r = 0.01, p = 0.96). The average daily dosage of delirium-related medications on ECMO days did not differ significantly. On an exploratory multivariable logistic regression, the proportion of delirium days was not associated with mortality. Conclusion: Longer duration of delirium was associated with lighter sedation and shorter paralysis, but it did not discern in-hospital mortality. Future studies should evaluate analgosedation and paralytic strategies to optimize delirium, sedation level, and outcomes.