Clinical and Experimental Emergency Medicine最新文献

Acute ischemic stroke (AIS) outcomes depend critically on rapid, accurate early diagnosis in the emergency department. Traditional prehospital tools and large vessel occlusion focused scales aid triage, but have limited ability to distinguish ischemic from hemorrhagic stroke, a distinction essential for treatment decision in acute phase. Recent advances include mobile stroke units equipped with computed tomography (CT), point-of-care labs, and telemedicine, as well as biomarkers, enabling field-based diagnosis and faster therapy initiation. In-hospital imaging strategies using CT, CT perfusion, and magnetic resonance imaging (MRI)-based tissue clocks have expanded eligibility for endovascular thrombectomy to large core and unclear onset wake-up strokes. Prolonged cardiac monitoring and highresolution vessel wall MRI enhance identification of embolic sources and high-risk atherosclerotic plaques. Artificial intelligence now supports rapid imaging interpretation, workflow optimization, and treatment selection. Tenecteplase, a novel thrombolytic, offers a practical alternative to alteplase with similar safety and efficacy, while post-thrombectomy care emphasizes individualized blood pressure management. In minor stroke or high-risk transient ischemic attack, short-term dual antiplatelet therapy reduces early recurrence, and early initiation of lipid lowering agents after AIS may stabilize unstable plaques and improve vascular outcomes. Together, these innovations mark a shift toward integrated, time-sensitive, and precision-based AIS care from prehospital assessment, emergency department care and post-reperfusion management.

Objective: Sepsis remains a major clinical challenge because of its complex, heterogeneous, and multidimensional clustering patterns. This study aimed to investigate the association between vasopressor administration and machine learning-derived clusters based on initial vital signs and lactate measurements obtained in emergency department (ED) and intensive care unit (ICU) settings.

Methods: A retrospective cohort analysis was performed using data from the Korean Shock Society Septic Shock (KOSS) Registry (septic shock in the ED) and the Marketplace for Medical Information in Intensive Care (MIMIC)-IV database (ICU patients with suspected infection). To derive clusters, k-means clustering was applied to six initial vital signs and serum lactate measurements. The primary outcome was vasopressor administration. Secondary outcomes included second vasopressor administration and 28-day mortality.

Results: A total of 17,500 patients were included in the analysis (KOSS cohort, n=7,130; MIMIC-IV cohort, n=10,370). K-means clustering identified three distinct clusters in each cohort. In the KOSS cohort, Cluster 3 was characterized by the lowest mean arterial pressure (MAP) (62 mmHg [IQR, 53-71]) and the highest diastolic shock index (DSI) (2.6 [2.3-3.0]). This cluster was associated with the highest rates of vasopressor administration (93.9%), second vasopressor administration (33.5%), and 28-day mortality (25.3%) (all p<0.001). Comparable physiological and clinical patterns were observed in the MIMIC-IV cohort, in which Cluster 3 likewise demonstrated the lowest MAP (68 mmHg [60-76]) and highest DSI (2.0 [1.8-2.3]). This group similarly exhibited the poorest outcomes, including vasopressor administration (41.0%), second vasopressor administration (16.7%), and 28-day mortality (29.0%).

Conclusion: Machine learning-derived clusters based on initial vital signs and serum lactate levels demonstrated different patterns of vasopressor use and mortality. The clinical utility of this approach for guiding timely or targeted vasopressor therapy requires prospective validation.

Objective: The diagnosis of acute ischemic stroke (AIS) is time-sensitive and reliant on neuroimaging, which is not always immediately accessible. This systematic review aims to identify and evaluate bloodbased biomarkers with potential to support early diagnostic decision-making and facilitate prompt referral for confirmatory imaging and treatment.

Methods: Following the PRISMA guidelines, a systematic search of PubMed, Web of Science, LILACS, Medline, Scopus, Scielo, Epistemonikos and TRIP Data Base was conducted for primary studies published in the last five years. Original studies evaluating blood-based biomarkers collected within 12 hours of symptom onset for early diagnosis of AIS were included. The methodological quality of included studies was assessed using the QUADAS-2 tool.

Results: Ten studies met the inclusion criteria, investigating a range of biomarkers including proteins, non-coding RNAs, and lipids. Most studies were case-control in design, with overall risk of bias rated as low to moderate. Multi-marker panels combining biomarkers with clinical scales (e.g., D-dimer and GFAP with FAST-ED; AUC = 0.95), and lipidomics-based models (AUC = 0.968), demonstrated the highest diagnostic performance. Several individual non-coding RNAs also showed promising accuracy (AUC > 0.85).

Conclusion: Blood-based biomarkers, especially when used in multi-marker panels, demonstrate considerable potential as triage tools for early AIS diagnosis. Their application in point-of-care settings could reduce diagnostic uncertainty and accelerate time to treatment. However, prospective validation in real-world emergency environments is essential prior to clinical implementation.

Objective: Early recognition of out-of-hospital cardiac arrest is frequently hindered by agonal or abnormal breathing, leading to delayed dispatcher-assisted cardiopulmonary resuscitation. Improving recognition of respiratory anomalies can optimize early intervention and survival. This scoping review aimed to map strategies used by emergency medical dispatchers to identify respiratory anomalies predictive of cardiac arrest during emergency calls.

Methods: A scoping review of the literature was performed following Joanna Briggs Institute recommendations and reported according to the PRISMA-ScR guidelines. The population, concept, and context (PCC) framework focused on (P) dispatchers and the (C) strategies to recognize breathing abnormalities predictive of cardiac arrest, (C) during emergency calls. Experimental, observational, and qualitative studies from PubMed, Scopus, Cochrane Library, and CINAHL up to June 30, 2025, were included.

Results: From 109 records, 16 studies met the inclusion criteria. Three thematic areas were identified: modifications to dispatch protocols, adding focused questions to improve detection of agonal breathing and cardiac arrest; hand-on-belly assessment, although rarely used, enhanced diagnostic accuracy but introduced modest delays in cardiopulmonary resuscitation initiation; and caller-reported breathing descriptors reduced missed or delayed recognition of cardiac arrest. The importance of dispatcher training, caller-related factors, and the use of technological aids were identified as secondary cross cutting themes.

Conclusions: Strategies such as protocol adjustments, hand-on-belly assessment, and systematic use of breathing descriptors may strengthen dispatcher recognition of cardiac arrest. Structured training on technical and communication skills is essential to optimize their application. Further research should evaluate the impact of these interventions on survival outcomes across diverse cultural and organizational contexts.

Background: Rescuer posture, position, and patient height during chest compressions (CC) influence its efficacy and efficiency. No clear recommendations exist on these aspects. It is essential to systematize the existing knowledge, especially for nurses and healthcare providers involved in resuscitation.

Purpose: To conduct a systematic review about the impact of rescuer posture, position, and manikin/patient height on CC efficacy and efficiency.

Methods: The study followed PRISMA guidelines and was registered on PROSPERO. Eligibility criteria included peer-reviewed articles or conference papers comparing different rescuer postures, positions, or manikin/patient heights during CC performed with both hands, regarding efficacy or efficiency. Databases consulted: MEDLINE Complete, SPORTDiscus, Cochrane Reviews, and CINAHL Complete. Methodological quality was assessed using the Quality Assessment Tool for Quantitative Studies.

Results: Of 6539 articles, only 34 met inclusion criteria. All were observational, used manikins, and were classified as weak in global methodological quality. Compared with standing, several studies suggested the kneeling posture may be associated with more effective and efficient CC. Evidence regarding the optimal patient height, including potential anthropometric-based adjustments, remains limited. Findings across the included studies indicated that variations in hand position appeared to have minimal influence on CC quality. Some studies reported decreased CC quality when rescuers performed CC while walking.

Conclusions: The available evidence suggested that performing CC while kneeling on a firm surface may be beneficial when feasible. Future research is needed to further evaluate the impact of bed height, self-selected rescuer position, and their relevance to emergency practice.

Objective: This study systematically reviews the causes, effects, and potential solutions to emergency department (ED) crowding, with emphasis on challenges amplified by the COVID-19 pandemic.

Methods: Following PRISMA guidelines, we searched MEDLINE, CINAHL, and Web of Science for peer-reviewed studies published from January 1, 2018, to January 31, 2025, that investigated ED crowding. Studies were included if they evaluated crowding causes, consequences, or interventions, using metrics such as ED length of stay, boarding, or left without being seen. Four reviewers independently screened titles, abstracts, and full texts. Study quality was assessed using the Scottish Intercollegiate Guidelines Network (SIGN) critical appraisal tools. This review was registered with PROSPERO (CRD420251117676).

Results: Of 23,408 studies identified, 226 met inclusion criteria. Most studies were retrospective (83%) and of low (62%) or acceptable (35%) quality. Crowding was primarily driven by input (high patient volumes, limited primary care access), throughput (staffing shortages, laboratory and imaging delays), and output (boarding, late discharges) factors. Adverse effects included increased mortality, treatment delays, prolonged inpatient stays, higher rates of patients leaving without being seen, and reduced patient satisfaction. Effective strategies included provider-in-triage, nurse-initiated orders, and split-flow models. Output-focused interventions, such as active bed management and early discharge protocols, required system wide coordination. The COVID-19 pandemic shifted patient volumes and led to innovative solutions such as drive-through clinics and repurposed spaces to alleviate surges.

Conclusion: ED crowding is a persistent global issue with significant clinical and operational consequences. While promising interventions exist, high-quality evidence remains limited, underscoring the need for system-level and multifaceted solutions.

Objective: We aimed to compare the 72-hour survival of the endotracheal intubation (ETI) with video laryngoscope (VL), ETI with direct laryngoscope (DL), and supraglottic airway (SGA) in out-of-hospital cardiac arrest (OHCA) patients in Korea.

Methods: This study included adult OHCA patients who received advanced airway management by designated response teams for severe disease, using a nationwide OHCA registry in South Korea from July 2019 to December 2021. The primary outcome was 72-hour survival, and secondary outcomes were survival to hospital discharge and good neurological recovery. Multivariable logistic regression was used, adjusted for confounders, to compare the outcomes among the three airway management methods.

Results: Among 77,629 OHCA cases, 10,857 were included. SGA was attempted in 9,379 cases, ETI with DL in 493 cases, and ETI with VL in 985 cases. The rates of any prehospital ROSC and 72-hour survival were 13.3% and 11.0% for SGA, 16.0% and 11.4% for ETI with DL, and 18.2% and 11.9% for ETI with VL. Compared to SGA, ETI with VL was significantly associated with 72-hour survival: adjusted odds ratio (OR) [95% confidence interval (CI)] 1.34 (1.06-1.70) for ETI with VL and 1.13 (0.81-1.56) for ETI with DL). There was no significant association between the type of AAM and survival to discharge or good neurological recovery.

Conclusion: In an emergency medical service system staffed by advanced emergency medical technician-level providers, ETI with VL might be associated with improved 72-hour survival compared to SGA. However, this short-term benefit did not extend to survival to hospital discharge.

Objective: Sepsis, a life-threatening organ dysfunction, remains a major global health concern. Early detection remains challenging due to nonspecific symptoms. Non-invasive tools such as the Shock Index (SI), Diastolic Shock Index (DSI), Capillary Refill Time (CRT), and Mottling Score (MS) may help assess hemodynamic status and predict mortality, but a comprehensive comparison of their prognostic value is lacking. This study compares the performance of these tools in predicting mortality at 24 hours, 7 days, and 28 days in septic patients.

Methods: A monocentric, prospective observational study was conducted from January to September 2024. Adult patients (≥18 years) with suspected infection and a National Early Warning Score-2 ≥5 were enrolled. Demographic data, vital signs, CRT, MS, and mortality outcomes were recorded at 24 hours, 7 days, and 28 days.

Results: A total of 135 patients were included (median age 85 years, IQR 79-90; 44.4% female). Mortality rates were 15.6% at 24 hours, 25.2% at 7 days, and 35.6% at 28 days. CRT showed the highest predictive value for 24-hour mortality (AUC: 0.829, p<0.01), while MS had the best performance at 7 days (AUC: 0.732, p<0.01) and 28 days (AUC: 0.749, p<0.01). No significant differences emerged in pairwise comparisons.

Conclusion: While no tool was found to significantly outperform others, CRT and MS provide valuable, non-invasive mortality prediction in sepsis. Indeed, CRT is most effective for early risk stratification, while MS correlates with mid- and long-term outcomes, supporting their integration into clinical assessment.

Objective: To develop and validate MEDIVAL, a progressive Chain-of-Thought (CoT) evaluation framework for automated assessment of large language model (LLM)-generated emergency department documentation that aligns with expert clinical judgment in acute care settings.

Methods: We developed a three-tier evaluation framework using Persona-based, Error-enhanced, and Insight-integrated strategies. The framework was tested across four LLMs (GPT-4o, GPT-4.1, Claude- 3.5, Claude-3.7) on 33 emergency department records evaluated by four expert emergency physicians. Each model assessed documents using three progressive CoT strategies across five criteria: Appropriateness, Accuracy, Structure/Format, Conciseness, and Clinical Validity. Evaluations were compared using Spearman's correlation, with differences assessed via Friedman test and Wilcoxon signed-rank test with Bonferroni correction. Reproducibility was evaluated using intraclass correlation coefficient (ICC) analysis.

Results: All models showed improved expert alignment as CoT complexity increased, with Claude-3.7 (r=0.712, P<0.001) and GPT-4o (r=0.702, P<0.001) achieving strongest correlations using Insightintegrated strategy. GPT-4.1 demonstrated largest relative improvement (43.3% increase from r=0.457 to r=0.655, P<0.001). Significant differences were found across strategies (χ²(2)=48.39, P<0.001), although Error-enhanced and Insight-integrated approaches showed a small but statistically significant difference (P=0.002). High reproducibility was confirmed (ICC > 0.919), with Claude-3.5 showing highest consistency (0.997-0.998).

Conclusions: MEDIVAL demonstrates that progressive CoT strategies systematically enhance automated evaluation of emergency department documentation while maintaining high reproducibility. This offers a viable pre-screening tool for reducing expert workload while supporting reliable AI integration into emergency medicine workflows.

Objective: Emergency physicians have become hesitant to treat patients with opioids because of downstream sequelae related to opioid use disorder. We enrolled a prospective cohort to determine whether a patient's experience receiving an IV opioid was associated with multiple opioid prescriptions. Specifically, we tested whether greater improvement in pain and a larger euphoric response could predict which previously opioid-naïve patients exposed to IV opioids would fill > 2 opioid prescriptions in the subsequent six-month period.

Methods: We determined 0-10 pain scores before and 15 minutes after opioid-naïve ED patients were administered IV opioids for treatment of severe pain. We also determined opioid-induced euphoria using 0-10 scales querying how good, how high, and how much euphoria the opioid caused and how likely the participant was to want the opioid again. Six-month outcomes were ascertained using the state prescription monitoring database.

Results: Among 506 patients, 32 (6.3%) filled two or more prescriptions during the six months following the ED visit. There were no differences between those who filled >2 prescriptions and those who did not with regard to pain relief (p= 0.54), how good the medication made participants feel (p=0.91), how high the medication made participants feel (p=0.97), how much euphoria the opioid caused (p=0.23), or how likely the participant was to want the medication again (p=0.37).

Conclusion: Filling >2 opioid prescriptions was uncommon after initial exposure to therapeutic IV opioids and was unrelated to either analgesic efficacy or opioid-induced euphoria.