Sunday 13 October

Sunday 13 October

Added to your list of favorites
Deleted from your list of favorites
14:35 - 14:55

ePoster 1.1 - Short Oral Presentation - Screen 1

Moderator: Barbra BACKUS (Emergency Physician) (Rotterdam, The Netherlands)
14:40 - 14:45 #18325 - SP002 Diagnostic value of Coronary Artery Disease Consortium and updated Diamond–Forrester scores in predicting obstructive coronary artery disease among emergency department patients with chest pain.
SP002 Diagnostic value of Coronary Artery Disease Consortium and updated Diamond–Forrester scores in predicting obstructive coronary artery disease among emergency department patients with chest pain.

Objective Assessing the pretest probability of coronary artery disease (CAD) is crucial for patients with chest pain at an emergency department (ED). Current guidelines recommend the use of the updated Diamond–Forrester (DF) method and Coronary Artery Disease (CAD) Consortium models by the American College of Cardiology/Amercian Heart Association (ACC/AHA) and European Society of Cardiology (ESC), respectively. In the situation which methods or models have not been proved definitely superior over others for assessing it, we studied to compare the performance of these models among patients with chest pain presented in the ED.

Methods We compared three scores (DF, CAD consortium basic, and clinical) among 536 patients with chest pain who underwent coronary computed tomographic angiography (CTA). Invasive angiography was performed to confirm the stenosis shown on CTA, if clinically indicated. Primary outcome was the presence of obstructive CAD (≧50% stenosis).

Results Overall, 174 (32.5%) patients were diagnosed with obstructive CAD. CAD consortium clinical model underestimated the prevalence of CAD (26.1%), and on the other hand, DF model overestimated (53.1%). To predict obstructive CAD, the CAD consortium clinical model had superior area under the receiver-operating curve (0.754), followed by the CAD consortium basic (0.736), and finally, the DF model (0.718). Whereas the CAD consortium models more accurately classified patients without any CAD or nonobstructive CAD as low-risk patients, the DF model more accurately classified high-risk patients with obstructive CAD. The net reclassification improvement of CAD consortium basic and clinical models were 15.3% and 18.0%, respectively.

Conclusion Compared with the DF model, the CAD consortium clinical model appears to be superior for the prediction of low-risk patients with <15% probability of having obstructive CAD. However, this model needs caution when using in high-risk population.

Yosep SHIN (SEOUL, Korea, Republic of), Ahn SHIN, Un Woo LEE
14:45 - 14:50 #18822 - SP003 Multilevel Prediction model of acute myocardial infarction in chest pain patients of emergency department.
SP003 Multilevel Prediction model of acute myocardial infarction in chest pain patients of emergency department.


Prediction models have been developed for acute myocardial infarction (AMI) among chest pain patients in the emergency department (ED). But usually, they included many variables or high-sensitive cardiac biomarkers which make them less optimal for generalization. We developed 3-level (prehospital, ED triage, ED doctor’s initial exam) prediction model that could be used in many emergency medical systems.


Multivariable logistic regression model (LR) and gradient boosting model (GBM) were developed on data from 8,673 ED visit for chest pain. Only variables which would be available shortly after patient presentation were used. 3-level modeling have been done and variables obtained in each level were chosen. Electrocardiogram (ECG) and high-sensitive cardiac biomarker were excluded for their fundamental diagnostic value. We evaluated performance by area under receiver operating characteristic curve (AUROC). Developed models were validated on validation data of 1,767 ED visit.


About 8,673 subjects, patients diagnosed as AMI in ED were 866. AMI patients more likely to be older male, show higher triage severity, use emergency medical service and present typical chest pain. Variables chosen for 3-level were as follows: age, sex, time from symptom onset, mental status, ambulance use in prehospital level, triage result, shock state, tachycardia or bradycardia in ED triage level and atypical presentation of chest pain, hemoglobin, glutamic oxaloacetic transaminase (AST), glutamic-pyruvic transaminase (ALT), abnormal chest radiograph result for ED doctor’s initial exam level. We calculated AUROC of LR/GBM model: 0.697/0.703 in prehospital, 0.731/0.732 in ED triage and 0.773/0.787 in ED doctor’s initial exam level.

In validation data of 1,767 patients, AMI patients were 222. We applied same prediction model developed in test data and AUROC were as follows: 0.695/0.704 in prehospital, 0.724/0.725 in ED triage and 0.784/0.788 in ED doctor’s initial exam level.

Discussion & Conclusions
We developed multi-level prediction model of AMI for chest pain patient who visit ED. GBM models showed slightly better performance in both data.

Dr Kihong KIM (Seoul, Korea, Republic of), Jeong Ho PARK, Young Sun RHO

Sunday 13 October

Added to your list of favorites
Deleted from your list of favorites
14:35 - 14:55

ePoster 1.2 - Short Oral Presentation - Screen 2

Moderator: Jochen BERGS (Hasselt, Belgium)
14:35 - 14:55 #18338 - SP005 The “Stroke Code” model in a low-resources first-level Emergency Department.
SP005 The “Stroke Code” model in a low-resources first-level Emergency Department.

Background: Systemic thrombolysis (STL) protocols are currently used in stroke centres around the world to reduce the treatment delay. These models were designed and most applied in tertiary hospitals with wide availability of neurology physicians and a dedicated stroke team. We brought  the American Heart Association/American Stroke Association (AHA/ASA) “Target: Stroke” initiative guidelines in a first-level Emergency Department (ED), where acute stroke patients are entirely managed by Emergency Physicians (EPs) with only radiology specialists consultants.


Methods: In 2018 a “Stroke Code” model was implemented in the ED of the Santa Maria Nuova first-level Hospital in Florence, including key components of the Helsinki model such as EP, radiologyst and laboratory technicians allert since ambulance transport, rapid Triage protocol, rapid EP evaluation, creation of a “stroke bag” with every stroke tool in use, early tissue plasminogen activator (Alteplase, tPA) preparation and infusion in the contrast tomography (CT) area, and prompt data feedback. A phone-alert protocol was established with our Territoral Emergency Service (TES): the EP of the TES operative base remotely identify the stroke codes from the ambulance team information and alert the ED. ER nurse apply a dedicate Triage protocol, characerized by the fast confirmation of ongoing deficit and acute (<4.5h) onset and immediate transfert in the emergency room (ER). EPs collect essential anamnestic and clinical data to identify indications and controindications to STL, while the ER nurses obtain at least one large-bore peripheric venous access, blood test samples  and check vital parameters before going to the CT-area. After the non contrast head CT-scan the radiology consultant give immediately a temporary answer reporting only the presence or absence of intracranial bleedings and/or radiologic STL contraindications, in order to allow the EP to start STL. After tPA bolus the patients undergo a contrast CT-scan. The 1-hour tPA infusion is beginned back in the ER. When a large-vessel obstruction is detected, a neuroradiological consultation is performed, and if indicated the patient is trasfered to local stroke hub centre for local treatment. Everyone involved in the stroke-code pathway had a specific 2-hour theorical training, implemented with one high-fidelity in-site simulation. Door-to-needle time (DTN) is registered, and all stroke patients data are collected in a dedicated registry to identify and correct specific delays and monitor activity.

Results: Before the stroke-model application the DTN [mean±standard deviation, (interquartile range)] was 76±33 (56-99) minutes. In the first year of stroke code model implementation the DTN was significantly lower 38±26 (20-50) minutes, p<0.001. Patient who received STL were more likely to experience a complete neurological recovery (59% treated vs 41% non-treated, p=0.002). Three (6%) patient had haemorragic complications, all of them had an acute stroke final diagnosis, and no one had permanent sequaele (mRs 0, 0 and 1, respectively).

Conclusion: The stroke-code model can be successfully applied after a brief training period  in a low-resources first-level ED in order to reduce DTN for acute ischaemic stroke treatment. In our experience the stroke-model performing was associated with a better functional outcome without any clinically relevant harm.

Dr Simone BIANCHI (Firenze, Italy), Francesco PROSPERI IOVI, Alessandra GIUELLO, Gabriele BANDINELLI, Federico LISI, Chiara ALAMANNI, Michele LANIGRA, Rita MARINO, Angela KONZE, Roberto CARPI, Vieri VANNUCCHI, Fererico MORONI, Giancarlo LANDINI
14:35 - 14:55 #18341 - SP006 Is LP useful in suspected SAH?
SP006 Is LP useful in suspected SAH?

Utility of lumbar puncture (LP) after a normal brain computed tomography (CT) scan in patients presenting to the emergency department with suspected Subarachnoid Haemorrhage: a retrospective cohort study


The diagnostic approach for patients presenting to EDs with severe, sudden-onset headache suspected for SAH remains challenging. 

Modern third generation computed tomography is shown to be extremely

sensitive in identifying subarachnoid haemorrhage when it is carried out

within six hours of headache onset and interpreted by a qualified radiologist, therefore may eliminate the need for lumbar puncture. However, some clinicians still perform LP even after a normal CTB even within this time frame, which is an invasive, time consuming procedures with known complications (i.e. post LP headache, Infection and bleeding at the site, Transient or permanent Neurological complications etc). 



Objective: To assess the utility of LP in emergency patients being evaluated for

possible subarachnoid haemorrhage after a negative non-contrast brain CT scan



We conducted a retrospective data analysis in three emergency departments in Monash Health in Victoria, Australia, focussing on patients presenting with concerning headache and being evaluated for possible subarachnoid haemorrhage between June 2013 and June 2018. Patients were excluded if they had a history of recurrent headaches or were discharged without further investigations. A diagnosis of Subarachnoid haemorrhage was defined by any of subarachnoid blood on computed tomography, xanthochromia in cerebrospinal fluid, or any red blood cells in final tube of cerebrospinal fluid collected with positive results on cerebral angiography.


Of 5746 enrolled eligible cases, 2039 (35.5%) were further investigated with a CT brain based on history and examination. 397 patients (19.5%) were diagnosed with SAH after CT, while 1642 (80.5%) had a negative CT scan. Of this remaining cohort, 388 (23.5%) patients underwent LP, and neither of them demonstrated a true positive SAH.

The 1254 patients with a negative CT scan whom did not undergo a lumbar puncture were followed up for 6 months by hospital and community record review. In this cohort of patients, 401 cases were lost to follow-up owing to a lack of available data. Of those followed-up, 1 patient died from haemorrhage stroke during his third hospitalisation.




LP is not required in all patients with suspected subarachnoid haemorrhage with a negative CT scan and it should be decided on a case by case basis. Further analysis is to determine if there are defining characteristics that eliminate the need for LP which can result in unnecessary risks and complications, with minimal benefit in diagnosing SAH.


Dr Pourya POURYAHYA (Melbourne, Australia), Alastair MEYER, Easaw-Mamutil NAOMI, Yan TAN, Hui (Glen) TENG
14:35 - 14:55 #18911 - SP007 A retrospective cohort study analyzing whether the ottawa ankle rules are correctly applied in the emergency department of a university hospital.
SP007 A retrospective cohort study analyzing whether the ottawa ankle rules are correctly applied in the emergency department of a university hospital.



The Ottawa Ankle Rules (OAR) were developed in 1992 in order to develop decision rules for the use of radiography in the emergency department (ED) and reduce unnecessary imaging. The purpose of this study is to evaluate how these OAR were applied in the ED of a university hospital between the 1st of July and the end of December of the year 2016. It examines how the OAR application in the ED has evolved compared to a similar study with data from 2015. The previous study led to changes in protocol as well as the implementation of training to improve the application of the OAR in triage. This study evaluates the success of these changes and trainings as well as whether predictive factors can be identified that drive the application of the OAR in the ED.


In a retrospective cohort study, a sample of patient records aged between 6 and 98 years old with ankle trauma were analyzed, using descriptive statistics, for the application of the OAR at triage and for use of imaging. Logistic regression was used to identify predictive factors.


The OAR were applied at triage in 90% of the cases. This is up from the 60% established by a previous, similar study a year before. However, imaging was still taken in 60% of cases where the OAR were negative. The study could identify some statistically significant predictive factors, but their predictive power is low as they explain 10% of the variation in the data of the OAR application.


The application of the OAR at triage has achieved a high level in this university hospital. The progress from 60% in 2015 to 90% now can be attributed to the training of medical staff in OAR application. The high percentage of imaging taken in OAR negative cases remains an area for further research to identify root causes.

Pauline SCHAUMANS (Jette, Belgium), Carol CLINCKAERT, Ives HUBLOUE
14:35 - 14:55 #18954 - SP008 Screening for pulmonary embolism in patients with chest pain in the emergency department: a retrospective multicenter study.
SP008 Screening for pulmonary embolism in patients with chest pain in the emergency department: a retrospective multicenter study.

Introduction Chest pain is a common chief complaint in the emergency department. Among its associated differential diagnoses, pulmonary embolism (PE) remains a key concern for the clinician. There are no clear recommendations on which patients should undergo a formal work-up for PE diagnosis. The purpose of our study was to determine the percentage of chest pain patients who are investigated for PE diagnosis and to determine the clinical profile of these patients. 

Methods We performed a retrospective multicenter study in 3 French Emergency Departments. We included all patients who visited these centers for a chest pain during a two month period. Patients were excluded if they were already treated for, or diagnosed with a thrombo-embolic event at the time of ED presentation. The primary outcome was the initiation of PE workup. This was defined by any evidence or mention in the chart of this diagnosis and reason for its rule-out or confirmation, namely order of D-dimers, CTPA, V/Q scan or lower limbs venous Doppler ultrasound. We also aimed to find factors associated to this outcome. 

Results We included 881 patients with a chief complaint of chest pain. Mean age was 49 years and 481 (56%) were men. A total of 263 patients (30%, 95% confidence interval 27% to 33%) underwent a formal PE workup, 235 (89%) of them had a D-dimer testing and 50 underwent a CTPA. Four patients had a lower limb venous doppler, and PE was explicitly ruled out on the basis of a PERC score of zero in 22 (8%) patients. PE was ultimately diagnosed in 7 cases (prevalence of 2.6%, 95% confidence interval 1.1% to 5.3%). In the multivariate logistic regression model, five factors were identified as independently associated with a workup for PE diagnosis: female gender, young age, no ischemic heart disease, recent flight, and associated dyspnea. 

Conclusion Among patients visiting the emergency department with a chest pain, 30% underwent work up for PE. We report five clinical variables independently associated with a higher probability of PE workup in our sample.

Antoine LEFEVRE-SCELLES, Dr Melanie ROUSSEL (Rouen), Paul JEANMAIRE, Anne-Laure PHILIPPON, Luc-Marie JOLY, Yonathan FREUND

Sunday 13 October

Added to your list of favorites
Deleted from your list of favorites
14:35 - 14:55

ePoster 1.3 - Short Oral Presentation - Screen 3

Moderator: Youri YORDANOV (Médecin) (Paris, France)
14:35 - 14:40 #18297 - SP009 Comparing of CPR related trauma: mechanical chest devices vs. manual CPR.
SP009 Comparing of CPR related trauma: mechanical chest devices vs. manual CPR.


Resuscitation (CPR) with mechanical chest devices are not recomanded for routine CPR according to randomised trials. One of possible explanation could be CPR related trauma caused with mechanical chest devices. Current data are based on subanalysis from randomised trial, but autopsies are limitated by law and autopsy results are not objectivised.


To compare injuries after CPR in autopsy results by manual resuscitated and mechanical (LUCAS 2, AutoPulse, CORpulse) resucitated patients and establish possible proportion of CPR related injuries on death without respect to cause of cardiac arrest.


Retrospective multicentric study based on autopsy reports by patients died after CPR, patients with traumatic cause of cardiac arrest were excluded. Patients were devided in two groups: mechanical and manual CPR. For objective evaluation of injury seriousness we used Abbreviated injury scale scoring for the most serious injury and New Injury Scale Score for summary of all injuries.


We have enroled 704 patients, after trauma exclusion we have analyzed 630 autopsies. Manual CPR were provided by 559 patients and mechanical by 64 patients. Both groups are no diferent in age, gender, bystander CPR anad cardiac etiology of Arrest. Mechanical CRP was significantly longer (p=0,0005). Both groups have no diferences in incidency of injuries of thoracic vessels, lungs, heart, pericard, pleura, stomach, liver and spleen. We have observed injuries by 80% of manual and 87,5% of mechanical CPR. The most frequent was thorax sceleton injury 85,5% vs. 87,5%. Median of the most seriuos injury was 3 ( serious by Abbreviated injury scale scoring) without statistical difference, median of summary of injuries (New Injury Severity Score) was 13 in both groups ( low probability of fatal injury). If we analysed CPR by LUCAS 2 compared to manual, results are similar, only pericard injuries are higher with LUCAS 2.


Incidency a seriousness of CPR related injuries according to autopsy reports are no diferent in comapring of manual and mechanical CPR. Mechanical CPR is significant longer a LUCAS 2 leads to significant more pericard injuries without influence to total seriousness of injury.

Jiri KARASEK (Prague, Czech Republic), Betka BLANKOVA, Andrea DOUBKOVA, Tereza PITASOVA, David NAHALKA, Tomas BARTES, Jiri HLADIK, Tomas ADAMEK
14:40 - 14:45 #18371 - SP010 Correlation between serum levels of lactate dehydrogenase and neurological outcomes in patients who undergo target temperature management after cardiac arrest.
SP010 Correlation between serum levels of lactate dehydrogenase and neurological outcomes in patients who undergo target temperature management after cardiac arrest.

Background: The optimal time to measure serum lactate dehydrogenase level (SLL) to predict prognosis in cardiac arrest (CA) survivors has not been elucidated. We aimed to compare the relationships between time-related SLL and neurological prognosis in CA survivors.

Methods: We conducted a retrospective study examining patients with CA who were treated with target temperature management (TTM). SLL was checked repeatedly at 24-h intervals after return of spontaneous circulation (ROSC). SLL at ROSC and 24-, 48-, and 72-h outcomes were the relationships between each time interval SLL and the neurological outcome 3 months post-CA.

Results: A total of 256 comatose patients with CA were treated with TTM. Seventy-three patients were included, and 31 patients (42%) experienced a good neurological outcome. At 24, 48, and 72 h, there was a significant difference between good and poor outcome groups (p<0.001), except at ROSC (p = 0.056). The area under the receiver operating curve (AUC) of at ROSC was 0.631 (95% confidence interval [CI], 0.502–0.761). The AUC at 48 h (0.830; 95% CI, 0.736 – 0.924) was higher than that at 24 and 72 h (0.786; 95% CI, 0.681–0.892 and 0.821; 95% CI, 0.724–0.919).

 Discussion & Conclusions: A higher SLL was strongly associated with and seemed predictive of poor outcomes. Furthermore, at 48 and 72 h, SLL may be a useful predictor of poor neurological outcomes. Prospective studies should be conducted to confirm these results.

Jin Hong MIN, Dr Hoil KIM (Daejeon, Korea, Republic of)
14:45 - 14:50 #18574 - SP011 The prognostic value of pre-arrest neutrophil-to-lymphocyte ratio for in-hospital mortality in adult patients with in-hospital cardiac arrest.
SP011 The prognostic value of pre-arrest neutrophil-to-lymphocyte ratio for in-hospital mortality in adult patients with in-hospital cardiac arrest.


  As an indicator of systemic inflammation, the neutrophil-to-lymphocyte ratio (NLR) has been proved to be associated with a prognosis of a range of inflammation-related diseases. Although the study found that post-arrest NLR can predict the poor outcomes in patients with in-hospital cardiac arrest (IHCA), the relationship between pre-NLR and worse prognostic of IHCA patients was unclear. This study aimed to investigate the association between pre-arrest NLR and in-hospital mortality in patients of IHCA. We hypothesized that pre-arrest NLR is related to in-hospital mortality of IHCA patients.


  This was a single-center retrospective cohort study recruited IHCA patients in the emergency department (ED) of West China hospital of Sichuan University between January 2016 and May 2017. This hospital is a 4300-bed tertiary teaching hospital and is one of the largest medical centers in the southwest of China. Consecutive patients with cardiac arrest in the ED were included in this study. We excluded patients younger than 18 years, major trauma, lack of necessary data for analysis and hematological diseases or receiving any treatment which might affect the pre-arrest NLR values. Patients were divided into two groups according to the outcomes of in-hospital mortality. Clinical information and blood sample results were collected. Multivariate regression models were used to evaluate the associations between pre-arrest NLR and in-hospital mortality. The receiver operating characteristic (ROC) curve was used to assess the predictive value of pre-NLR.


  Out of 488 IHCA patients during the study period a total of 385 (78.89%) patients were eligible for analysis, of those 267(69.35%) were male and mean age was 60.63±17.27. Of 385 patients, 64 (16.62%) survived to discharge. Patients with in-hospital mortality had a significantly higher pre-arrest NLR compared with survival to discharge patients (11.32[6.98,17.68] vs. 3.65[3.16,6.01], p<0.001). In the univariate model, pre-arrest NLR was associated with in-hospital mortality (OR: 1.347, 95% CI: 1.222-1.484, p<0.001). In the multivariate adjustment, higher pre-arrest NLR was independently associated with in-hospital mortality (AOR=1.276, 95%CI:1.160-1.403, p<0.001) after adjusting for age, gender, history of renal insufficiency, total CPR duration, globulin, alanine transaminase and aspartate aminotransferase.Furthermore, the prognostic performance of pre-arrest NLR was excellent (AUC: 0.86 [95%CI: 0.80-0.92, p<0.001]).


  In this retrospective observational study, we found that the excellent predictive ability of pre-NLR to predict in-hospital mortality for patients resuscitated from IHCA. We demonstrated the pre-arrest NLR is also a predictor for in-hospital mortality in IHCA patients. Therefore, we have reason to speculate that the systemic inflammatory response and the potential immune dysfunction before resuscitation in critically ill patients are associated with poor prognosis after resuscitation. Timely and effective medical interventions for critically ill patients might improve the survival when IHCA occurs.


  Pre-arrest NLR is a useful predictor of in-hospital mortality in adults with IHCA.  

Ethical approval and informed consent

   The study was conducted in line with the Declaration of Helsinki and gained approval by the Ethical Committee of West China Hospital of Sichuan University(Reference number: 2019201).

Dr Sheng YE (ChengDu, China), Junzhao LIU, Yarong HE, Pr Yu CAO


Knowledge of Basic Life Support (BLS) techniques generates an undisputed benefit by improving survival prognosis in any CPR case, provided that BLS measures are initiated within the first 4 minutes of the CPR, and the comprehensive emergency system included within the "Chain of Survival" is implemented. International experience has shown that learning the instrumental management of the airway by the "first responders" is useful for increasing survival rates. For early defibrillation to be possible, knowledge about the use of semi-automatic defibrillation needs to be widely disseminated among staff in different health units.


To know the degree of knowledge in Cardiopulmonary Resuscitation (CPR) among the health personnel of 10 urban and rural health centers prior to the realization of a plan of fromación in the form of workshops of eminently practical content.


Prior to the workshop, 120 surveys were distributed among the participants consisting of 20 questions with 2 or 3 answers in Likert format referring to demographic data (profession, age and sex) and different aspects of basic resuscitation techniques with instrumental support and drugs. The surveys were analyzed with the SPSS version 20 database.


The response rate was 100%. The degree of knowledge of the different techniques is shown in the tables provided, although in general terms, a knowledge deficit is detected in all the techniques analysed, which increases as the complexity of these techniques increases. This is despite the fact that 94.9% of the professionals admit to having received a course in CPR and 89.8% are aware of the ABCDE alert system.


The level of knowledge in cardiopulmonary resuscitation techniques demonstrated in this study by health professionals working in Primary Care is scarce, which makes it very necessary for health authorities to implement training programmes if we want to improve the response of these professionals to cardiac arrests in the hospital setting.


Sunday 13 October

Added to your list of favorites
Deleted from your list of favorites
14:35 - 14:55

ePoster 1.4 - Short Oral Presentation - Screen 4

Moderator: Dr Thomas BEATTIE (Senior lecturer) (Edinburgh)
14:35 - 14:40 #18200 - SP013 Salivary biomarkers role in pain diagnostics.
SP013 Salivary biomarkers role in pain diagnostics.


The key to adequate pain management is assessing its presence and identifying exact severity of the pain. Current ‘gold-standard’ pain assessment tools rely on self-reporting, requiring an ability to communicate this personal experience. Self-reporting varies from patient to patient and could be inaccurately understood by healthcare professionals. According to the study results, acute as well as chronic pain remains one of the most misunderstood, under-diagnosed, and under-treated medical problems, particularly in children. Pain diagnosis and management would benefit from the development of objective markers of nociception and pain.


To investigate concentration of salivary cortisol and melatonin in children with acute pain and compare it with severity of pain and changes in vital signs.


We conducted a pilot observational study in Lithuanian University of Health Sciences Hospital Kauno Klinikos Pediatric emergency department (PED). Twenty six patients complaining of acute pain referred to PED were included into the study. Patients having chronical conditions (cancer, immunodeficiency, diabetes etc.), fever, dehydration or chronic pain were excluded. We recorded patient’s gender, age, vital signs (heart rate (HR), blood pressure (BP), respiratory rate (RR), temperature (t°) and oxygen saturation (SaO2)), pain characteristics (severity and duration of pain according the used pain scale and its localization). Saliva samples were collected and were stored in -80°C till analysis was performed. Samples were analyzed using cortisol and melatonin ELISA kits.


Sixteen boys and 10 girls were involved in our research. Age median was10 (4-16) years. Fourteen cases were trauma patients, 12 cases referred due to pain of other origin then trauma. Analyzing vital signs, we noticed HR and BP increase with regard to pain. Other parameters (RR, t°, SaO2) were within the age range. The median of cortisol and melatonin levels were 287.5 (68-1330) pg/ml and 17,6 (8,6-46,8) pg/ml respectively. There were several findings related to saliva hormone level and intensity of pain, duration of pain and it’s link to vital signs There was a tendency to melatonin reduction with increased intensity of pain (p=0,136). The longer the pain lasted, the higher cortisol levels were identified (p=0.01). However, there was no link between abnormal vital signs and changes in our biomarkers.


Our primary results show a cortisol rise with regard to pain in time dependent manner. Melatonin levels decreased in relation to increased pain intensity. These results show a potential of cortisol and melatonin as biomarkers in acute pain diagnostics. 

Kristina GANZIJEVA (Kaunas, Lithuania, Lithuania), Lina JANKAUSKAITE
14:40 - 14:45 #18203 - SP014 Association of Neutrophil to Lymphocyte Ratio and Platelet to Lymphocyte Ratio with In-Hospital Mortality in Patients with Gastrointestinal Perforation Undergoing Surgery.
SP014 Association of Neutrophil to Lymphocyte Ratio and Platelet to Lymphocyte Ratio with In-Hospital Mortality in Patients with Gastrointestinal Perforation Undergoing Surgery.

Objective: The Neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) are recognized markers of inflammation associated with poor outcomes in various clinical situations. Gastrointestinal perforation (GIP) is a life-threatening disease with a high mortality rate. We analyzed the prognostic significance of NLR and PLR in patients with gastrointestinal perforation (GIP) undergoing surgery.

Methods: This was a multi-center observational retrospective study. We reviewed electronic medical records of adult patients with GIP admitted to three academic hospitals between January 2009 and December 2018, who also received surgical operation. We obtained demographic and clinical data of GIP patients. Multivariate logistic regression model was used to determine the predictive value of NLR and PLR on in-hospital mortality and to evaluated risk factors associated with in-hospital mortality. The primary outcome was all-cause in-hospital mortality.

Results: Among 9279 patients, 879 adult patients with GIP underwent surgical operation. Seventy eight patients (8.9%) were died and 801 (91.1%) were survived. In the Multivariate logistic regression analysis, factors associated with in-hospital mortality were female, underlying chronic renal failure, C-reactive protein >100mg/l and Albumin <3.5g/dl (Adjusted odds ratio [95%CI]; 2.73 [1.07-6.97], 4.20 [1.83-9.68], 8.43 [2.29-31.03], 5.36 [2.29-12.59], respectively).

Conclusion: NLR and PLR are not associated with mortality in patients with gastrointestinal perforation undergoing surgery in the study. Female, underlying chronic renal failure, C-reactive protein>100 mg/l and Albumin<3.5 g/dl may help to identify high-risk patients.
Yeonghoon BAE (Seoul, Korea, Republic of)
14:45 - 14:50 #19059 - SP015 Utility of s-tryptase in Emergency Department patients with possible anaphylaxis.
SP015 Utility of s-tryptase in Emergency Department patients with possible anaphylaxis.

Background: Diagnosing anaphylaxis is straight forward in typical cases but can be challenging if the presentation is atypical. Generally it requires that patients have acute symptoms from more than one of the following systems; skin or mucosa, respiratory, cardiovascular or gastrointestinal. In addition to clinical diagnosis it has been shown that in patients with vague or atypical symptoms considered possibly due to an acute allergic reaction, evaluating s-tryptase can give additional diagnostic information, ideally if done within 30-180 min from onset of symptoms. Measuring s-tryptase also has the potential to diagnose mastocytosis, a rare but probably underdiagnosed condition of mast cell proliferation.

In our emergency department (ED), staff has been educated since 2011 on obtaining a s-tryptase level in cases where a patient could possibly be having an acute allergic reaction but a definitive diagnosis cannot be made based on clinical evaluation. Further evaluation by an allergist during an outpatient follow up is also recommended. The aim of this study was to assess how useful obtaining a s-tryptase level was on the work up of patients with possible anaphylaxis in the ED.

Methods: With institutional review board approval, all cases where a s-tryptase level was obtained from ED patients during the period from 2011-2018 were retrospectively reviewed. A database was collected including information on patient demographics, presenting symptoms and signs, treatment given, diagnosis, s-tryptase level and follow up.

Results: During the study period a total of 214 patients in the ED had s-tryptase measured. Females were 131 (61.2%) and average age 40.6 years (range 11-88). When evaluating patients, 60 (28.0%) of patients had only symptoms from one organ system, 70(32.7%) from two, 55(25.7%) from three and 26(12.1%) of the patients had symptoms from four organ systems. Three patients (1.4%) did not have symptoms from any of the four target organ systems.  Of the patients, 86.4% had skin or mucosal symptoms, 47.7% cardiovascular symptoms, 49.5% respiratory symptoms and 36.0% had gastrointestinal symptoms. Blood was drawn for s-tryptase analaysis within the recommended time frame in 133(62.1%) of the cases. Serum tryptase was elevated (>12µg/l) in 36 (16.8%) cases.

Of the 214 cases, 126 returned for further evaluation by an allergist and 65 (51.6%) of those were considered to have had an episode of anaphylaxis.

When evaluating the cases where blood samples for s-tryptase analysis had been collected within the recommended time frame of 30-180 minutes and returned for further evaluation by an allergist, the sensitivity of s-tryptase to diagnose anaphylaxis was 40,91% (95% CI 26,34% - 56,75%) and specificity 96,30% (95% CI 81,03% - 99,91%). No case of mastocytosis was identified in the patient cohort.

Discussion and conclusions: Obtaining a s-tryptase level on ED patients with possible anaphylaxis is specific but not sensitive in diagnosing atypical cases of anaphylaxis. No cases of mastocytosis were identified in the patient cohort suggesting that mastocytosis is uncommon among ED patients in our community.

Karolina HANSEN, Hjalti Mar BJORNSSON (Reykjavik, Iceland), Maria I GUNNBJORNSDOTTIR
14:50 - 14:55 #19219 - SP016 High cortisol levels predict worse outcome in patients with community pneumonia.
SP016 High cortisol levels predict worse outcome in patients with community pneumonia.

Critical Illness-Related Corticosteroid Insufficiency (CIRCI) is a condition still not completely understood. Although the current guidelines suggest the administration of corticosteroids only in septic patients who do not respond to the initial fluid resuscitation, it is not clear how it works and the signaling pathways that may affect its efficacy.

Therefore, we made the hypothesis that corticosteroids therapeutic action may be related not only to cortisol levels, but also to the degree of glucocorticoids receptors (GR) expression.

We included in this study 181 patients presented to the Emergency Room with the diagnosis of community pneumonia. Blood samples were collected at the admission and outcomes were evaluated during hospital stay and after 30 days.

Cytokines were measured by Multiplex method, and glucocorticoid receptors α and β were measured in leukocytes homogenates by ELISA. Data are presented here as mean±SEM. 

Patients’ ages vary from 18.2 to 100.0 y.o (average = 61.7±1.5). From the 181 patients, 96 were also septic and 40 had septic shock. Mortality was 17.7% (32 patients). Seventeen of these patients were chronic corticoid users and were evaluated separately. 

The main discriminators between survivors and non survivors were SOFA (2.9±0.2 vs 4.8±0.4, respectively, p<0.05) and total cortisol levels (18.5±2.1 vs 39.7±8.2, respectively, p<0.05). There was no difference in GR α or β expression between survivors and non survivors. Previous corticosteroid use also did not affect the levels of GRs expression, suggesting that these receptors are not subjected to feedback regulation.

In addition none of the inflammatory markers measured at admission (proadrenomedulin, procalcitonin, C-reactive protein or cytokines) were able to distinguish the patients who would survive.  

In conclusion, we showed that cortisol levels are a good predictor of outcome in patients with community pneumonia and GR expression did not affect this outcome. 

Lucas MARINO (Sao Paulo, Brazil), Hermes BARBEIRO, Denise BARBEIRO, Luzmarina GOMEZ, Julio ALENCAR, Julio MARCHINI, Rodrigo BRANDAO, Heraldo SOUZA

Sunday 13 October

Added to your list of favorites
Deleted from your list of favorites
14:35 - 14:55

ePoster 1.5 - Short Oral Presentation - Screen 5

Moderator: Isabel LUECK (Resident) (Hamburg, Germany)
14:35 - 14:40 #18284 - SP017 Temporal evolution of the PCVCO2-PaCO2 / CaO2-CcvO2 ratio versus serum lactate during resuscitation in septic patients with hypotension.
SP017 Temporal evolution of the PCVCO2-PaCO2 / CaO2-CcvO2 ratio versus serum lactate during resuscitation in septic patients with hypotension.

Bakground: Lactate as a target for resuscitation in patients with septic shock has important limitations. The PcvCO2-PaCO2 / CaO2-CcvO2 ratio may be used as an alternative for the same.The primary objective of the study was to evaluate the correlation between serum lactate and PcvCO2-PaCO2 / CaO2-CcvO2 ratio measured at various time points to a maximum of 24h in patients with septic shock (Mean arterial pressure < 65mmHg). The secondary objectives were to study the  1) Relationship between the PcvCO2-PaCO2 / CaO2-CcvO2 ratio and lactate clearance at 6, 12 and 24 hrs as compared to the initial serum lactate. 2) To study the correlation between the arterial lactate and the PcvCO2-PaCO2 gap at each measurement. 3) Outcome in terms of ICU length of stay, organ dysfunction and mortality at day 28.

Methods: This prospective observational cohort study was conducted at the mixed ICU(Medical/Surgical)   of the All India Institute of Medical Sciences (AIIMS), New Delhi from July,2016 to April,2018.Thirty patients with sepsis-induced hypotension (MAP < 65mmHg) who were being actively resuscitated were enrolled. Paired arterial and central venous blood samples were obtained 0.5 hourly till stabilisation of MAP (maximum of two samples), and 6 hourly thereafter for the first 24h. Patients were followed up to day 28 of enrolment for mortality and organ system failure. All statistical analysis was performed using ‘Stata’ software (Ver 15.1; StataCorp LLC, Texas, USA). Correlation between the various variables was done using the Spearman coefficient. Subgroup analysis of variables between survivors and non-survival groups was done using the Wilcoxon-Mann-Whitney test. Sensitivity and specificity of the PcvCO2-PaCO2)/(CaO2-CcvO2) ratio and arterial lactate were calculated and Receiver-Operating-Characteristic curves were constructed.

Results: A positive correlation was observed between arterial lactate and  PcvCO2-PaCO2 / CaO2-CcvO2 ratio at 0h, 6h, 12h, 18h (R=0.413 P=0.02; R=0.567 P=0.001; R=0.408 P=0.025; R=0.521 P=0.003, respectively). No correlation was seen between  PcvCO2-PaCO2 / CaO2-CcvO2 ratio and  lactate clearance. The subgroup analysis showed that neither an abnormal arterial lactate (> 2mmol/L) nor an abnormal PcvCO2-PaCO2 / CaO2-CcvO2 ratio (>1) at the time of enrolment could distinguish survivors from non-survivors (at D28 of enrolment). The median (PcvCO2-PaCO2)/(CaCO2-CcvO2) ratio was higher in non-survivors than in survivors at all time points. However, this reached statistical significance only at the 24h time point9 P=0.004). A PcvCO2-PaCO2 / CaO2-CcvO2 ratio>1.696 at 24h of resuscitation predicted mortality at 28d (Sensitivity: 80%, Specificity 69.2%, AU-ROC 0.82). This threshold also distinguished survival at D28 in the Kaplan Meier estimates (Chi-square=6.00, P=0.014). An arterial lactate > 1.6mmol/L at 24h of resuscitation predicted mortality at 28d (Sensitivity 73.33%, Specificity 69.23%, AU-ROC 0.853). This threshold also distinguished survival at D28 in the Kaplan Meier estimates (Chi-square=5.62, P=0.018)

Discussion and Conclusion(s): The (PcvCO2-PaCO2)/(CaCO2-CcvO2) ratio and the lactate are positively correlated during the first 24h following active resuscitation from sepsis-induced hypotension. The (PcvCO2-PaCO2)/(CaCO2-CcvO2) ratio at 24h is significantly higher in non-survivors, and a threshold of 1.696 mmHg/mL/dL for (PcvCO2-PaCO2)/(CaCO2-CcvO2) ratio at 24h significantly differentiates survivors from non-survivors. The (PcvCO2-PaCO2)/(CaCO2-CcvO2) ratio may be used as an end-point of haemodynamic resuscitation from septic shock.

Pr Vimi REWARI (New Delhi, India), Shyam MADABHUSHI, Anjan TRIKHA, Rahul Kumar ANAND, Rashmi RAMACHANDRAN
14:40 - 14:45 #18475 - SP018 P25/30 somatosensory evoked potential is superior to N20 in predicting neurological recovery after cardiac arrest: A prospective, observational study.
SP018 P25/30 somatosensory evoked potential is superior to N20 in predicting neurological recovery after cardiac arrest: A prospective, observational study.


The absence of N20 somatosensory evoked potential (SEP) is regarded as a predictor of poor outcome after cardiac arrest with very high specificity. However, its sensitivity in predicting the poor outcome is unsatisfactory. The P25/30 SEP is a positive deflection following N20 with a latency of 25-35 msec. According to our prior study, N20 without following P25/30 is related to poor outcome, while N20 followed by P25/30 is highly related to good outcome. In this study, we evaluated whether the amplitude of P25/30 predicts neurologic recovery after cardiac arrest more accurately than the amplitude of N20.


This is a prospective multicenter observational study. Patients were consecutively enrolled in four university-affiliated teaching hospitals. SEPs of comatose survivors after out-of-hospital cardiac arrest treated by 33℃-targeted temperature management were recorded 72 hours after return of spontaneous circulation. The cutoff value of P25/30 and N20 amplitude showing 100% specificity in predicting poor neurological recovery was determined by receiver operating characteristic (ROC) analysis. We defined poor neurological recovery as the worst Cerebral Performance Category score higher than three during the admission period. We performed penalized maximum likelihood estimation in logistic regression analysis. Thereafter, we compared the area under curve (AUC) of the P25/30-based model predicting poor outcome to AUC of the N20-based model. According to the previous study, a total of 86 subjects would be required to detect a 0.05 difference in AUC with a power of 80% and a type I error of 5%. Values of p less than 0.05 were considered statistically significant.


Out of a total of 87 patients included in the study, 43 patients showed good neurological recovery, while 44 patients showed poor neurological recovery. The cutoff values of SEP amplitudes showing 100% specificity in predicting poor neurological recovery were 0.63 μV (P25/30) and 0 μV (N20). Sensitivity in predicting poor neurological recovery of P25/30 was 0.86 (95% confidence interval 0.73 - 0.95), while N20 was 0.7 (95% confidence interval 0.55 - 0.83). In addition to N20 or P25/30, cardiac arrest rhythm and anoxic time were selected as independent variables for the multivariable logistic regression models. The AUC of the P25/30-based model was 0.958 (95% confidence interval 0.92 - 1), while the AUC of the N20-based model was 0.911 (95% confidence interval 0.85 - 0.98). AUC comparison between the N20-based model and the P25/30-based model showed a statistically significant difference (p = 0.02).


P25/30 showed superior value in predicting poor neurologic recovery after cardiac arrest than the N20. P25/30 showed higher sensitivity in predicting poor neurologic recovery than the N20, which implies potential as a predictor of good neurologic recovery.

Sang Hoon OH, Pr Joo Suk OH (Seoul, Korea, Republic of), Hyunho JUNG, Jungtaek PARK, Ji Hoon KIM, Jeong Ho PARK, Jung Hee WEE, Seung Pill CHOI, Kyu Nam PARK
14:45 - 14:50 #18956 - SP019 Eligibility for organ donation: Is there a place in emergency department ?
SP019 Eligibility for organ donation: Is there a place in emergency department ?

Introduction: Actually, 60% of organ procurement is perfomed in patients who died of brain death due to a severe stroke, mainly haemorrhagic stroke (75%). The objective of this study is to identify patients in emergency departments for severe stroke who may be eligible for organ retrieval procedures. 

Methods: We realized an epidemiological, descriptive, observational, monocentric, retrospective study at a university hospital emergency department for one year including all strokes with a Glasgow coma scale ≤ 12 and/or a NIHSS score ≥ 17. We considered patients who died early (death < 3 days) and with no neoplasia as potentially eligible for organ donation and compared the characteristics of eligible patients to patients whose brain death has been confirmed.

Results: 1582 patients were hospitalized for stroke, 312 patients had a severe stroke. 201 patients, with severe stroke, was not managed in intensive care. In this cohort according to predefined criteria, 34 patients were considered eligible for multi-organ procurement. Fifteen (44%) had an ischemic stroke and 19 (56%) had a hemorrhagic stroke. The initial mean Glasgow coma scale was 6 (± 3). The median age was 84 years [77-89]. We compared patients with confirmed brain death (N=32) to potentially eligible donors (N=34). Age was significantly higher in potentially eligible patients (65 years vs 84 years; p < 0.0001). There was no significant difference in the Glasgow coma scale for antiaggregant or anticoagulant treatment between the 2 groups. 13 patients were cared for and died in neurology, 15 patients in the emergency department and 6 patients in short-term hospital units. Among potentially eligible patients, 75% of patients died within 48 hours. For patients with confirmed brain death, 75% of patients died within 72 hours. 

Conclusion: The identification of patients who may be eligible for organ retrieval is difficult in the emergency department. A better awareness of practitioners could make it possible to better filiarize the management of these patients.

Dr Melanie ROUSSEL (Rouen), Antoine LEFEVRE-SCELLES, Edgar MENGUY, Mathieu BENHADDOUR, Luc-Marie JOLY
14:50 - 14:55 #19102 - SP020 Analysis of Single Substance Heroin Exposures reported to the U.S. Poison Centers from Healthcare Facilities.
SP020 Analysis of Single Substance Heroin Exposures reported to the U.S. Poison Centers from Healthcare Facilities.

Background: Heroin use has reached a public health crisis in the U.S. Since 2010, the rate for deaths involving heroin has almost tripled, from 1.5 per 100,000 in 2011 to 5.1 in 2016. The number of people using heroin for the first time in the U.S. has increased in recent years. Hence it is important to track heroin overdoses, especially those reported from the healthcare facilities (HCF) as these may greatly increase resource use. The objective of the current study is to use outline the epidemiology of single substance heroin exposures reported to the National Poison Data System from the HCFs.

Methods: The NPDS was queried for all human single substance exposures to heroin reported to the U.S. Poison Centers (PCs) from HCFs between 2011 and 2017. We descriptively assessed the relevant demographic and clinical characteristics. Trends in heroin frequencies and rates (per 100,000 human exposures from HCF) were analyzed using Poisson regression methods. Percent changes were reported with the corresponding 95% confidence intervals (95% CI).

Results: There were 15,692 single substance heroin exposures reported to the PCs from HCFs. The number of calls increased from 1,142 to 3,865 during the study period. Among these calls, 90.2% were reported from acute care hospitals and emergency departments (EDs), 6.5% were reported from freestanding EDs, while 3.3% were reported from physician offices.  Acute exposures to heroin were responsible for 67.7% of the calls from HCF. Approximately 19% of the patients reporting such heroin exposures were admitted to the critical care unit (CCU), with 56.6% patients treated and released. Residence was the most common site of exposure (69.7%). Among the patients, m were male, with the majority of the individuals between ages 20 and 39 years (70.9%). Pediatric cases accounted for 6.3% of the exposures. Intentional abuse (74.5%) and misuse (10.5%) were commonly observed reasons for exposure. During the study period, the proportion of heroin abuse cases increased (73.3% to 76.1%). Major effects were seen in 19.1% cases and the mortality rate for single substance heroin exposures from HCF was 1.8%. Notably, the number of heroin-related fatalities in this group doubled during the study period.  Coma (26.8%) and respiratory depression (27.3%) were frequently observed clinical effects. Naloxone (60.7%) was the most frequently reported therapy. During the study period, the frequency of heroin exposures increased by 238.4% (95% CI: 215.8%, 261.5%; p<0.001), and the rate of heroin exposures increased by 187.6% (95% CI: 149.8%, 231.2%; p<0.001).

Discussion: There was a significant increase in single substance heroin exposures reported to the PCs from HCFs during the study period. This increase may be a result of the lower cost of heroin and the tighter regulations on the prescribing of opioids. Changes in the sources of supply and potency of heroin products can result in substantial adverse events seen in the HCFs. Exposures reported to the poison centers further highlight the need for sustained, targeted, and multifactorial responses to the ongoing opioid epidemic, including timely surveillance.

Saumitra REGE (Charlottesville, VA, USA), Dr Christopher HOLSTEGE