Handling of derived imbalanced dataset using XGBoost for identification of pulmonary embolism—a non-cardiac cause of cardiac arrest

Author(s):  
Naira Firdous ◽  
Sushil Bhardwaj
Circulation ◽  
2018 ◽  
Vol 138 (Suppl_2) ◽  
Author(s):  
Shinichi Ijuin ◽  
Akihiko Inoue ◽  
Nobuaki Igarashi ◽  
Shigenari Matsuyama ◽  
Tetsunori Kawase ◽  
...  

Introduction: We have reported previously a favorable neurological outcome by extracorporeal cardiopulmonary resuscitation (ECPR) for out of hospital cardiac arrest. However, effects of ECPR on patients with prolonged pulseless electrical activity (PEA) are unclear. We analyzed etiology of patients with favorable neurological outcomes after ECPR for PEA with witness. Methods: In this single center retrospective study, from January 2007 to May 2018, we identified 68 patients who underwent ECPR for PEA with witness. Of these, 13 patients (19%) had good neurological outcome at 1 month (Glasgow-Pittsburgh Cerebral Performance Category (CPC):1-2, Group G), and 55 patients (81%) had unfavorable neurological outcome (CPC:3-5, Group B). We compared courses of treatment and causes/places of arrests between two groups. Results are expressed as mean ± SD. Results: Patient characteristics were not different between the two groups. Time intervals from collapse to induction of V-A ECMO were also not significantly different (Group G; 46.1 ± 20.2 min vs Group B; 46.8 ± 21.7 min, p=0.92). Ten patients achieved favorable neurological outcome among 39 (26%) with non-cardiac etiology. In cardiac etiology, only 3 of 29 patients (9%) had a good outcome at 1 month (p=0.08). In particular, 5 patients of 10 pulmonary embolism, and 4 of 4 accidental hypothermia responded well to ECPR with a favorable neurological outcome. Additionally, 6 of 13 (46%), who had in hospital cardiac arrest, had good outcome, whereas 7 of 55 (15%) who had out of hospital cardiac arrest, had good outcome (p=0.02). Conclusions: In our small cohort of cardiac arrest patients with pulmonary embolism or accidental hypothermia and PEA with witness, EPCR contributed to favorable neurological outcomes at 1 month.


Circulation ◽  
2015 ◽  
Vol 132 (suppl_3) ◽  
Author(s):  
Rasmus Aagaard ◽  
Philip Caap ◽  
Nicolaj C Hansson ◽  
Morten T Bøtker ◽  
Asger Granfeldt ◽  
...  

Introduction: Survival from non-shockable cardiac arrest is unlikely unless a reversible cause is identified and treated. Guidelines state that ultrasound has the potential to identify reversible causes. Currently, ultrasonographic findings from patients with spontaneous circulation are extrapolated to patients in cardiac arrest. While right ventricular (RV) dilation is a finding normally associated with pulmonary embolism (PE), porcine studies have shown that RV dilation is also seen in ventricular fibrillation (VF) and severe hypoxia. No studies have investigated how causes of cardiac arrest affect RV size during resuscitation. Hypothesis: The RV diameter is larger during resuscitation of cardiac arrest caused by PE when compared to hypoxia and VF. Methods: Pigs were anesthetized and randomized to cardiac arrest induced by VF, hypoxia, or PE. Advanced life support (ALS) was preceded by 7 minutes of untreated cardiac arrest. Cardiac ultrasound images of the RV from a subcostal 5-chamber view were obtained during induction of cardiac arrest and ALS. The RV diameter was measured two centimeters from the aortic valve at end diastole. RV diameter at 3rd rhythm analysis was the primary endpoint. Based on pilot studies a sample size of 8 animals in each group was needed. Results: Eight animals were included in each group. RV diameter was not statistically different at baseline (mean (95%CI)) in VF: 19.8 (18.0-21.5) mm, hypoxia: 19.8 (16.6-22.9) mm, and PE: 21.8 (19.2-24.3) mm. During induction of cardiac arrest the RV diameter increased to 29.6 (27.3-31.9) mm in the hypoxia group and 38.0 (33.4-42.6) mm in the PE group (difference to baseline and between groups, both p<0.01). Induction of VF caused an immediate increase in the RV diameter to 25.0 (21.2-28.8) mm (difference to baseline p<0.01). At 3rd rhythm analysis, RV diameter was 32.4 (28.6-36.2) mm in the PE group, which was significantly larger than both the hypoxia group at 23.3 (19.5-27.0) mm and the VF group at 24.9 (22.2-27.5) mm (difference between groups p<0.01). Conclusions: Cardiac arrest due to VF, hypoxia, and PE all caused an increase in RV diameter. During resuscitation the RV was larger in PE compared to VF and hypoxia. Cardiac ultrasound thus has the potential to detect PE during resuscitation.


2015 ◽  
Vol 21 (2) ◽  
pp. 113-116
Author(s):  
Yeo-Ul Yun ◽  
Sang-Min Shim ◽  
Yun-Sook Kim

CHEST Journal ◽  
2021 ◽  
Vol 160 (4) ◽  
pp. A2201
Author(s):  
Chun Siu ◽  
Megan Fisher ◽  
Ashley Vojtek ◽  
Tyler Boozel

2020 ◽  
Vol 2020 ◽  
pp. 1-5
Author(s):  
Sumit Sohal ◽  
Akhilesh Thakur ◽  
Aleena Zia ◽  
Mina Sous ◽  
Daniela Trelles

Disseminated Intravascular Coagulation (DIC) is a disorder of coagulation which is commonly seen as a complication of infections, traumas, obstetric diseases, and cancers especially hematological and rarely solid cancers. DIC may rarely be the presenting feature of an undiagnosed malignancy. It may present in the form of different phenotypes which makes its diagnosis difficult and leads to high mortality. The treatment comprises supportive, symptomatic treatment and removal of the underlying source. Here, we present a patient with history of being on warfarin for atrial fibrillation and other comorbidities who presented with elevated INR of 6.3 and increasing dyspnea on exertion. Over the course of her stay, her platelet counts started dropping with a concurrent decrease in fibrinogen levels. She eventually developed pulmonary embolism, followed by stroke and limb ischemia, which was indicative of the thrombotic phenotype of DIC. Her pleural fluid analysis showed huge burden of malignant cells in glandular pattern suggestive of adenocarcinoma and was started on heparin drip. However, the patient had cardiac arrest and expired on the same day of diagnosis.


Sign in / Sign up

Export Citation Format

Share Document