Ventricular Remodeling Following Metabolic and Bariatric Surgery Decreases Need for Heart Transplantation: A Predictive Model.

PURPOSE: For patients with obesity and congestive heart failure (CHF) who require heart transplantation (HT), aggressive weight loss has been associated with ventricular remodeling, or subclinical alterations in left and right ventricular structure that affect systolic function. Many have suggested offering metabolic and bariatric surgery (MBS) for these patients. As such, we evaluated the role of MBS in HT for patients with obesity and CHF using predictive modelling techniques. MATERIALS AND METHODS: Markov decision analysis was performed to simulate the life expectancy of 30,000 patients with concomitant obesity, CHF, and 30% ejection fraction (EF) who were deemed ineligible to be waitlisted for HT unless they achieved a BMI < 35 kg/m2. Life expectancy following diet and exercise (DE), Roux-en-Y gastric bypass (RYGB), and sleeve gastrectomy (SG) was estimated. Base case patients were defined as having a pre-intervention BMI of 45 kg/m2. Sensitivity analysis of initial BMI was performed. RESULTS: RYGB patients had lower rates of HT and received HT quicker when needed. Base case patients who underwent RYGB gained 2.2 additional mean years survival compared with patients who underwent SG and 10.3 additional mean years survival compared with DE. SG patients gained 6.2 mean years of life compared with DE. CONCLUSION: In this simulation of 30,000 patients with obesity, CHF, and reduced EF, MBS was associated with improved survival by not only decreasing the need for transplantation due to improvements in EF, but also increasing access to HT when needed due to lower average BMI.

Prospective Evaluation of a Universally Applied Laparoscopic Gastric Ischemic Preconditioning Protocol Prior to Esophagectomy with Comparison with Historical Controls.

BACKGROUND: Anastomotic leak after esophagectomy is associated with significant morbidity and mortality. Our institution began performing laparoscopic gastric ischemic preconditioning (LGIP) with ligation of the left gastric and short gastric vessels prior to esophagectomy in all patients presenting with resectable esophageal cancer. We hypothesized that LGIP may decrease the incidence and severity of anastomotic leak. METHODS: Patients were prospectively evaluated following the universal application of LGIP prior to esophagectomy protocol in January 2021 until August 2022. Outcomes were compared with patients who underwent esophagectomy without LGIP from a prospectively maintained database from 2010 to 2020. RESULTS: We compared 42 patients who underwent LGIP followed by esophagectomy with 222 who underwent esophagectomy without LGIP. Age, sex, comorbidities, and clinical stage were similar between groups. Outpatient LGIP was generally well tolerated, with one patient experiencing prolonged gastroparesis. Median time from LGIP to esophagectomy was 31 days. Mean operative time and blood loss were not significantly different between groups. Patients who underwent LGIP were significantly less likely to develop an anastomotic leak following esophagectomy (7.1% vs. 20.7%, p = 0.038). This finding persisted on multivariate analysis [odds ratio (OR) 0.17, 95% confidence interval (CI) 0.03-0.42, p = 0.029]. The occurrence of any post-esophagectomy complication was similar between groups (40.5% vs. 46.0%, p = 0.514), but patients who underwent LGIP had shorter length of stay [10 (9-11) vs. 12 (9-15), p = 0.020]. CONCLUSIONS: LGIP prior to esophagectomy is associated with a decreased risk of anastomotic leak and length of hospital stay. Further, multi-institutional studies are warranted to confirm these findings.

Variability in Heart Yield From Donation After Brain Death Between Organ Procurement Organizations: An Opportunity for Improvement.

Increasing the number of available hearts for transplantation is the best strategy to decrease waitlist mortality. This study examines organ procurement organizations (OPOs) and their role in the transplantation network to determine whether variability in performance exists across them. Adult deceased donors who met the criteria for brain death between 2010 and 2020 (inclusive) in the United States were examined. A regression model was fitted and internally validated using donor characteristics available at the time of organ recovery to predict the likelihood of heart transplantation. Subsequently, an expected heart yield was calculated for each donor using this model. Observed-to-expected (O/E) heart yield ratios for each OPO were calculated by dividing the number of hearts recovered for transplantation by the expected number of recoveries. There were 58 OPOs active during the study period, and on average, OPO activity grew over time. The mean O/E ratio among OPOs was 0.98 (standard deviation ± 0.18). Twenty-one OPOs consistently performed below the expected level (95% confidence intervals < 1.0) and generated a deficit of 1,088 expected transplantations during the study period. The proportion of hearts that were recovered for transplantation varied significantly by OPO categories: low tier 31.8%, mid tier 35.6%, and high tier 36.2% (p < 0.01), even as the expected yield was similar across tiers (p = 0.69). OPO performance accounts for 28% of the variability in successfully transplanting a heart after accounting for the role of referring hospitals, donor families, and transplantation centers. In conclusion, there is significant variability in volume and heart yield from brain-dead donors across OPOs.

A proposed clinical coagulation score for research in trauma-induced coagulopathy.

BACKGROUND: Trauma-induced coagulopathy (TIC) has been the subject of intense study for greater than a century, and it is associated with high morbidity and mortality. The Trans-Agency Consortium for Trauma-Induced Coagulopathy, funded by the National Health Heart, Lung and Blood Institute, was tasked with developing a clinical TIC score, distinguishing between injury-induced bleeding from persistent bleeding due to TIC. We hypothesized that the Trans-Agency Consortium for Trauma-Induced Coagulopathy clinical TIC score would correlate with laboratory measures of coagulation, transfusion requirements, and mortality. METHODS: Trauma activation patients requiring a surgical procedure for hemostasis were scored in the operating room (OR) and in the first ICU day by the attending trauma surgeon. Conventional and viscoelastic (thrombelastography) coagulation assays, transfusion requirements, and mortality were correlated to the coagulation scores using the Cochran-Armitage trend test or linear regression for numerical variables. RESULTS: Increased OR TIC scores were significantly associated with abnormal conventional and viscoelastic measurements, including hyperfibrinolysis incidence, as well as with higher mortality and more frequent requirement for massive transfusion ( p < 0.0001 for all trends). Patients with OR TIC score greater than 3 were more than 31 times more likely to have an ICU TIC score greater than 3 (relative risk, 31.6; 95% confidence interval, 12.7-78.3; p < 0.0001). CONCLUSION: A clinically defined TIC score obtained in the OR reflected the requirement for massive transfusion and mortality in severely injured trauma patients and also correlated with abnormal coagulation assays. The OR TIC score should be validated in multicenter studies. LEVEL OF EVIDENCE: Prognostic and Epidemiological; Level IV.

Zone 1 Endovascular Balloon Occlusion of the Aorta vs Resuscitative Thoracotomy for Patient Resuscitation After Severe Hemorrhagic Shock.

Importance: Aortic occlusion (AO) is a lifesaving therapy for the treatment of severe traumatic hemorrhagic shock; however, there remains controversy whether AO should be accomplished via resuscitative thoracotomy (RT) or via endovascular balloon occlusion of the aorta (REBOA) in zone 1. Objective: To compare outcomes of AO via RT vs REBOA zone 1. Design, Setting, and Participants: This was a comparative effectiveness research study using a multicenter registry of postinjury AO from October 2013 to September 2021. AO via REBOA zone 1 (above celiac artery) was compared with RT performed in the emergency department of facilities experienced in both procedures and documented in the prospective multicenter Aortic Occlusion for Resuscitation in Trauma and Acute Care Surgery (AORTA) registry. Propensity score matching (PSM) with exact institution matching was used, in addition to subgroup multivariate analysis to control for confounders. The study setting included the ED, where AO via RT or REBOA was performed, and participants were adult trauma patients 16 years or older. Exposures: AO via REBOA zone 1 vs RT. Main Outcomes and Measures: The primary outcome was survival. Secondary outcomes were ventilation-free days (VFDs), intensive care unit (ICU)-free days, discharge Glasgow Coma Scale score, and Glasgow Outcome Score (GOS). Results: A total of 991 patients (median [IQR] age, 32 [25-48] years; 808 male individuals [81.9%]) with a median (IQR) Injury Severity Score of 29 (18-50) were included. Of the total participants, 306 (30.9%) had AO via REBOA zone 1, and 685 (69.1%) had AO via RT. PSM selected 112 comparable patients (56 pairs). REBOA zone 1 was associated with a statistically significant lower mortality compared with RT (78.6% [44] vs 92.9% [52]; P = .03). There were no significant differences in VFD greater than 0 (REBOA, 18.5% [10] vs RT, 7.1% [4]; P = .07), ICU-free days greater than 0 (REBOA, 18.2% [10] vs RT, 7.1% [4]; P = .08), or discharge GOS of 5 or more (REBOA, 7.5% [4] vs RT, 3.6% [2]; P = .38). Multivariate analysis confirmed the survival benefit of REBOA zone 1 after adjustment for significant confounders (relative risk [RR], 1.25; 95% CI, 1.15-1.36). In all subgroup analyses (cardiopulmonary resuscitation on arrival, traumatic brain injury, chest injury, pelvic injury, blunt/penetrating mechanism, systolic blood pressure ≤60 mm Hg on AO initiation), REBOA zone 1 offered an either similar or superior survival. Conclusions and Relevance: Results of this comparative effectiveness research suggest that REBOA zone 1 provided better or similar survival than RT for patients requiring AO postinjury. These findings provide the ethically necessary equipoise between these therapeutic approaches to allow the planning of a randomized controlled trial to establish the safety and effectiveness of REBOA zone 1 for AO in trauma resuscitation.

Hemorrhagic shock and tissue injury provoke distinct components of trauma-induced coagulopathy in a swine model.

INTRODUCTION: Tissue injury (TI) and hemorrhagic shock (HS) are the major contributors to trauma-induced coagulopathy (TIC). However, the individual contributions of these insults are difficult to discern clinically because they typically coexist. TI has been reported to release procoagulants, while HS has been associated with bleeding. We developed a large animal model to isolate TI and HS and characterize their individual mechanistic pathways. We hypothesized that while TI and HS are both drivers of TIC, they provoke different pathways; specifically, TI reduces time to clotting, whereas, HS decreases clot strength stimulates hyperfibrinolysis. METHODS: After induction of general anesthesia, 50 kg male, Yorkshire swine underwent isolated TI (bilateral muscle cutdown of quadriceps, bilateral femur fractures) or isolated HS (controlled bleeding to a base excess target of - 5 mmol/l) and observed for 240 min. Thrombelastography (TEG), calcium levels, thrombin activatable fibrinolysis inhibitor (TAFI), protein C, plasminogen activator inhibitor 1 (PAI-1), and plasminogen activator inhibitor 1/tissue-type plasminogen activator complex (PAI-1-tPA) were analyzed at pre-selected timepoints. Linear mixed models for repeated measures were used to compare results throughout the model. RESULTS: TI resulted in elevated histone release which peaked at 120 min (p = 0.02), and this was associated with reduced time to clot formation (R time) by 240 min (p = 0.006). HS decreased clot strength at time 30 min (p = 0.003), with a significant decline in calcium (p = 0.001). At study completion, HS animals had elevated PAI-1 (p = 0.01) and PAI-1-tPA (p = 0.04), showing a trend toward hyperfibrinolysis, while TI animals had suppressed fibrinolysis. Protein C, TAFI and skeletal myosin were not different among the groups. CONCLUSION: Isolated injury in animal models can help elucidate the mechanistic pathways leading to TIC. Our results suggest that isolated TI leads to early histone release and a hypercoagulable state, with suppressed fibrinolysis. In contrast, HS promotes poor clot strength and hyperfibrinolysis resulting in hypocoagulability.

Venovenous extracorporeal membrane oxygenation support in patients with COVID-19 respiratory failure: A multicenter study.

Objective: The COVID-19 pandemic presents a high mortality rate amongst patients who develop severe acute respiratory distress syndrome (ARDS). The purpose of this study was to evaluate the outcomes of venovenous extracorporeal membrane oxygenation (VV-ECMO) in COVID-19-related ARDS and identify the patients who benefit the most from this procedure. Methods: Adult patients with COVID-19 and severe ARDS requiring VV-ECMO support at 4 academic institutions between March and October 2020 were included. Data were collected through retrospective chart reviews. Bivariate and multivariable analyses were performed with the primary outcome of in-hospital mortality. Results: Fifty-one consecutive patients underwent VV-ECMO with a mean age of 50.4 years; 64.7% were men. Survival to hospital discharge was 62.8%. Median intensive care unit and hospitalization duration were 27.4 days (interquartile range [IQR], 17-37 days) and 34.5 days (IQR, 23-43 days), respectively. Survivors and nonsurvivors had a median ECMO cannulation time of 11 days (IQR, 8-18) and 17 days (IQR, 12-25 days). The average postdecannulation length of stay was 17.5 days (IQR, 12.4-25 days) for survivors and 0 days for nonsurvivors (IQR, 0-6 days). Only 1 nonsurvivor was able to be decannulated. Clinical characteristics associated with mortality between nonsurviors and survivors included increasing age (P = .0048), hemorrhagic stroke (P = .0014), and postoperative dialysis (P = .0013) were associated with mortality in a bivariate model and retained statistical significance in a multivariable model. Conclusions: This multicenter study confirms the effectiveness of VV-ECMO in selected critically ill patients with COVID-19-related severe ARDS. The survival of these patients is comparable to non-COVID-19-related ARDS.

A combat casualty relevant dismounted complex blast injury model in swine.

BACKGROUND: Improvised explosive devices have resulted in a unique polytrauma injury pattern termed dismounted complex blast injury (DCBI), which is frequent in the modern military theater. Dismounted complex blast injury is characterized by extremity amputations, junctional vascular injury, and blast traumatic brain injury (bTBI). We developed a combat casualty relevant DCBI swine model, which combines hemorrhagic shock (HS) and tissue injury (TI) with a bTBI, to study interventions in this unique and devastating military injury pattern. METHODS: A 50-kg male Yorkshire swine were randomized to the DCBI or SHAM group (instrumentation only). Those in the DCBI group were subjected to HS, TI, and bTBI. The blast injury was applied using a 55-psi shock tube wave. Tissue injury was created with bilateral open femur fractures. Hemorrhagic shock was induced by bleeding from femoral arteries to target pressure. A resuscitation protocol modified from the Tactical Combat Casualty Care guidelines simulated battlefield resuscitation for 240 minutes. RESULTS: Eight swine underwent the DCBI model and five were allocated to the SHAM group. In the DCBI model the mean base excess achieved at the end of the HS shock was -8.57 ± 5.13 mmol·L -1 . A significant coagulopathy was detected in the DCBI model as measured by prothrombin time (15.8 seconds DCBI vs. 12.86 seconds SHAM; p = 0.02) and thromboelastography maximum amplitude (68.5 mm DCBI vs. 78.3 mm in SHAM; p = 0.0003). For the DCBI models, intracranial pressure (ICP) increased by a mean of 13 mm Hg, reaching a final ICP of 24 ± 7.7 mm Hg. CONCLUSION: We created a reproducible large animal model to study the combined effects of severe HS, TI, and bTBI on coagulation and ICP in the setting of DCBI, with significant translational applications for the care of military warfighters. Within the 4-hour observational period, the swine developed a consistent coagulopathy with a concurrent brain injury evidenced by increasing ICP.

Succinate Activation of SUCNR1 Predisposes Severely Injured Patients to Neutrophil-mediated ARDS.

OBJECTIVES: Identify the metabolites that are increased in the plasma of severely injured patients that developed ARDS versus severely injured patients that did not, and assay if these increased metabolites prime pulmonary sequestration of neutrophils (PMNs) and induce pulmonary sequestration in an animal model of ARDS. We hypothesize that metabolic derangement due to advanced shock in critically injured patients leads to the PMNs, which serves as the first event in the ARDS. Summary of Background Data: Intracellular metabolites accumulate in the plasma of severely injured patients. METHODS: Untargeted metabolomics profiling of 67 critically injured patients was completed to establish a metabolic signature associated with ARDS development. Metabolites that significantly increased were assayed for PMN priming activity in vitro. The metabolites that primed PMNs were tested in a 2-event animal model of ARDS to identify a molecular link between circulating metabolites and clinical risk for ARDS. RESULTS: After controlling for confounders, 4 metabolites significantly increased: creatine, dehydroascorbate, fumarate, and succinate in trauma patients who developed ARDS ( P < 0.05). Succinate alone primed the PMN oxidase in vitro at physiologically relevant levels. Intravenous succinate-induced PMN sequestration in the lung, a first event, and followed by intravenous lipopolysaccharide, a second event, resulted in ARDS in vivo requiring PMNs. SUCNR1 inhibition abrogated PMN priming, PMN sequestration, and ARDS. Conclusion: Significant increases in plasma succinate post-injury may serve as the first event in ARDS. Targeted inhibition of the SUCNR1 may decrease ARDS development from other disease states to prevent ARDS globally.

Role of phosphodiesterase 1 in the pathophysiology of diseases and potential therapeutic opportunities.

Cyclic nucleotide phosphodiesterases (PDEs) are superfamily of enzymes that regulate the spatial and temporal relationship of second messenger signaling in the cellular system. Among the 11 different families of PDEs, phosphodiesterase 1 (PDE1) sub-family of enzymes hydrolyze both 3',5'-cyclic adenosine monophosphate (cAMP) and 3',5'-cyclic guanosine monophosphate (cGMP) in a mutually competitive manner. The catalytic activity of PDE1 is stimulated by their binding to Ca2+/calmodulin (CaM), resulting in the integration of Ca2+ and cyclic nucleotide-mediated signaling in various diseases. The PDE1 family includes three subtypes, PDE1A, PDE1B and PDE1C, which differ for their relative affinities for cAMP and cGMP. These isoforms are differentially expressed throughout the body, including the cardiovascular, central nervous system and other organs. Thus, PDE1 enzymes play a critical role in the pathophysiology of diseases through the fundamental regulation of cAMP and cGMP signaling. This comprehensive review provides the current research on PDE1 and its potential utility as a therapeutic target in diseases including the cardiovascular, pulmonary, metabolic, neurocognitive, renal, cancers and possibly others.

Strategies for successful implementation of resuscitative endovascular balloon occlusion of the aorta in an urban Level I trauma center.

BACKGROUND: The rationale for resuscitative endovascular balloon occlusion of the aorta (REBOA) is to control life-threatening subdiaphragmatic bleeding and facilitate resuscitation; however, incorporating this into the resuscitative practices of a trauma service remains challenging. The objective of this study is to describe the process of successful implementation of REBOA use in an academic urban Level I trauma center. All REBOA procedures from April 2014 through December 2019 were evaluated; REBOA was implemented after surgical faculty attended a required and internally developed Advanced Endovascular Strategies for Trauma Surgeons course. Success was defined by sustained early adoption rates. METHODS: An institutional protocol was published, and a REBOA supply cart was placed in the emergency department with posters attached to depict technical and procedural details. A focused professional practice evaluation was utilized for the first three REBOA procedures performed by each faculty member, leading to internal privileging. RESULTS: Resuscitative endovascular balloon occlusion of the aorta was performed in 97 patients by nine trauma surgeons, which is 1% of the total trauma admissions during this time. Each surgeon performed a median of 12 REBOAs (interquartile range, 5-14). Blunt (77/97, 81%) or penetrating abdominopelvic injuries (15/97, 15%) comprised the main injury mechanisms; 4% were placed for other reasons (4/97), including ruptured abdominal aortic aneurysms (n = 3) and preoperatively for a surgical oncologic resection (n = 1). Overall survival was 65% (63/97) with a steady early adoption trend that resulted in participation in a Department of Defense multicenter trial. CONCLUSION: Strategies for how departments adopt new procedures require clinical guidelines, a training program focused on competence, and a hospital education and privileging process for those acquiring new skills. LEVEL OF EVIDENCE: Therapeutic, level V.

Plasmin thrombelastography rapidly identifies trauma patients at risk for massive transfusion, mortality, and hyperfibrinolysis: A diagnostic tool to resolve an international debate on tranexamic acid?

BACKGROUND: Trauma patients with hyperfibrinolysis and depletion of fibrinolytic inhibitors (DFIs) measured by thrombelastography (TEG) gain clot strength with TXA, but TEG results take nearly an hour. We aimed to develop an assay, plasmin TEG (P-TEG), to more expeditiously stratify risk for massive transfusion (MT), mortality, and hyperfibrinolysis. METHODS: Trauma patients (N = 148) were assessed using TEG assays without exogenous additives (rapid/native), with exogenous plasmin (P-TEG) or tissue plasminogen activator (tPA TEG). The plasmin dose used does not effect healthy-control clot lysis 30 minutes after maximum amplitude (LY30) but causes shortened reaction time (R time) relative to native TEG (P-TEG R time < native TEG R time considered P-TEG negative). If P-TEG R time is greater than or equal to native TEG R time, the patient was considered P-TEG positive. Each assay's ability to predict MT, mortality, and (risk for) hyperfibrinolysis was determined. χ and Mann-Whitney U tests were used to compare categorical and continuous variables, respectively. Results were reported as median ± interquartile range or n (%). RESULTS: Plasmin TEG provided results faster than all other assays (4.7 ± 2.5-9.1 minutes), approximately 11-fold faster than rapid-TEG (rTEG) LY30 (54.2 ± 51.1-58.1 minutes; p < 0.001). Plasmin TEG-positive patients had greater than fourfold higher MT rate (30% vs. 7%; p = 0.0015) with an area under the receiver operating characteristic curve of 0.686 (p = 0.028), greater than fourfold higher 24-hour mortality (33.3% vs. 7.8%; p = 0.0177), greater than twofold higher 30-day mortality (35% vs. 16.4%; p = 0.0483), higher rates of DFI (55% vs. 18%; p < 0.001), and a trend toward elevated D-dimer (19.9 vs. 3.3 µg/mL; p = 0.14). Plasmin TEG was associated with hyperfibrinolysis on rTEG LY30 at the 7.6% threshold (p = 0.04) but not the 3% threshold (p = 0.40). Plasmin TEG performed best in relation to DFI, with a positive predictive value of 58% and negative predictive value of 81%. When combined with tPA TEG time to maximum amplitude, P-TEG outperformed rTEG LY30 for predicting MT (area under the receiver operating characteristic curve, 0.811 vs. 0.708). CONCLUSION: Within 5 minutes, P-TEG can stratify patients at highest risk for MT, mortality, and risk for hyperfibrinolysis. In composite with tPA TEG time to maximum amplitude, P-TEG outperforms rTEG LY30 for predicting MT and does so four times faster (12.7 vs. 54.1 minutes). The rapid results of P-TEG may be useful for those who practice selective TXA administration to maximize TXA's time-dependent efficacy. LEVEL OF EVIDENCE: Diagnostic test, level V.

Tranexamic acid is associated with reduced complement activation in trauma patients with hemorrhagic shock and hyperfibrinolysis on thromboelastography.

: Trauma with hemorrhagic shock causes massive tissue plasminogen activator release, plasmin generation, and hyperfibrinolysis. Tranexamic acid (TXA) has recently been used to treat bleeding in trauma by preventing plasmin generation to limit fibrinolysis. Trauma patients also have increased complement activation that correlates with mortality and organ failure, but the source of activation is not clear, and plasmin has recently been shown to efficiently cleave C3 and C5 to their activated fragments. We hypothesized that trauma patients in hemorrhagic shock with hyperfibrinolysis on thromboelastography (TEG) LY30 would have increased complement activation at early time points, as measured by soluble C5b-9 complex, and TXA would prevent this. Plasma samples were obtained from an unrelated, previously performed IRB-approved prospective randomized study of trauma patients. Three groups were studied with n = 5 patients in each group: patients without hyperfibrinolysis (TEG LY30 < 3%) (who therefore did not get TXA), patients with hyperfibrinolysis (TEG LY30 > 3%) who did not get TXA, and patients with hyperfibrinolysis who were then treated with TXA. We found that patients who did not receive TXA, regardless of fibrinolytic phenotype, had elevated soluble C5b-9 levels at 6 h relative to emergency department levels. In contrast, all five patients with initial TEG LY30 more than 3% and were then treated with TXA had reduced soluble C5b-9 levels at 6 h relative to emergency department levels. There were no differences in PF1 + 2, Bb, or C4d levels between groups, suggesting that coagulation and complement activation pathways may not be primarily responsible for the observed differences.

Clinical predictors of in-hospital mortality in venoarterial extracorporeal membrane oxygenation.

INTRODUCTION: Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is utilized as a life-saving procedure and bridge to myocardial recovery for patients in refractory cardiogenic shock. Despite technical advancements, VA-ECMO retains high mortality. This study aims to identify the clinical predictors of in-hospital mortality after VA-ECMO to improve risk stratification for this tenuous patient population. METHODS: The REgistry for Cardiogenic Shock: Utility and Efficacy of Device Therapy database is a multicenter, observational registry of ECMO patients. From 2013 to 2018, 789 patients underwent VA-ECMO. Bivariate analysis was performed on more than 300 variables regarding their association with in-hospital mortality. Logistic regression analyses were performed with variables chosen based upon clinical and statistical significance in the bivariate analysis. Tests were considered significant at a two-sided P < .05. RESULTS: Although 63.5% patients were successfully weaned from VA-ECMO, in-hospital mortality was 57.9%. Nonsurvivors were older (P < .0001), had higher body mass index (P = .01), higher rates of hypertension (P = .02), coronary artery disease (P = .02), chronic obstructive pulmonary disease (P = .02), chronic liver disease (P = .008), percutaneous coronary intervention (P = .02), and surgical revascularization (P = .02). Multivariate predictors for in-hospital mortality include older age (odds ratio [OR], 1.019; P = .007), cardiac arrest (OR, 2.76; P = .006), chronic liver disease (OR, 8.87; P = .04), elevated total bilirubin (OR, 1.093; P < .0001), and the presence of a left ventricular vent (OR, 2.018; P = .03). Pre-ECMO sinus rhythm was protective (OR, 0.374; P = .006). CONCLUSIONS: In a large study of recent VA-ECMO patients, in-hospital mortality remains significant, but acceptable given the severe pathology manifested in this population. Identification of pre-ECMO predictors of mortality helps stratify high-risk patients when deciding on ECMO placement, prolonged support, and prognosis.