Improving the overall standard of diagnosis and treatment for critically ill patients is a crucial component of the "Healthy China" strategy, directly impacting patient survival rates. This imperative underscores the need to refine our talent cultivation system in critical care medicine. It is essential to optimize pedagogical approaches, shifting the focus of education from mere "knowledge transmission" to comprehensive" competency development", ultimately enhancing the clinical competence of intensivists. Artificial intelligence offers unique advantages in this endeavor, particularly in optimizing teaching methodologies, enhancing clinical simulation training, and personalizing learning trajectories. As such, AI is poised to become a key direction for reshaping learning resources and learning strategies in critical care education.

Large language models (LLMs) are emerging as a transformative tool in critical care medicine, a discipline that fundamentally relies on multidisciplinary collaboration, real-time monitoring, and timely intervention. With their capability to integrate and synthesize multi-modal data (e.g., clinical notes, physiological signals), LLMs demonstrate significant potential to enhance condition monitoring, provide clinical decision support, and improve overall workflow efficiency. However, the rigorous demands of critical care -including the needs for individualized clinical reasoning, accurate interpretation of dynamic high-stakes patient data, and strict adherence to evidence - based practice-present substantial challenges for the direct application of general - purpose LLMs. This review systematically explores the potential application pathways for LLMs in critical care medicine by integrating insights from technical explorations and clinical practice. Furthermore, we propose a future roadmap focusing on four key areas: deep integration into core clinical workflows, continuous technological refinement, establishment of robust standards and validation frameworks, and fostering effective collaboration between clinicians and engineers.

The epidemiology of invasive pulmonary fungal disease (IPFD) has evolved during the coronavirus disease 2019 (COVID-19) pandemic, characterized by the expanded susceptible population and significant increase in incidence and mortality. COVID-19 is a key driver of this epidemiological shift, as the SARS-CoV-2 virus and its associated treatments can compromise normal immune defenses, thereby elevating the risk of pulmonary fungal infection. Early recognition of risk factors for IPFD and standardization of clinical practices are crucial for improving the timeliness of diagnosis and intervention. This article presents a systematic review of the research on IPFD during the COVID-19 pandemic, aiming to elucidate the evolving epidemiology and discuss the challenges in clinical diagnosis and treatment.

In recent years, critical care medicine, a core discipline dedicated to the management of critically ill patients, is increasingly recognized for its vital role in healthcare. Its advancement is pivotal to the overall capacity of the medical system and the outcomes of critically ill patients. In the context of the high incidence, disability rates, and mortality rates of severe conditions like sepsis, acute respiratory distress syndrome, and multiple organ failure, coupled with a growing demand for precision medicine, establishing high-level research platforms has become an imperative for propelling the field forward. This article discusses the critical role of such platforms in critical care medicine, elaborating on their significant contributions to basic research, translational science, multidisciplinary collaboration, and innovative technology development. Taking the Key Laboratory of Multiple Organ Failure of the Ministry of Education as an example, we propose key strategies for building high-level platforms, emphasizing the cultivation of a learning culture, talent development, sustainable funding, and the establishment of novel collaborative paradigms. Looking ahead, with the deepening integration of artificial intelligence and big data technologies, high-level research platforms are poised to drive critical care medicine towards greater intelligence and precision, providing robust support for enhancing treatment capabilities and achieving the strategic goals of the "Healthy China" initiative.

To evaluate the impact of a real-time critical care early warning platform on outcomes in patients with orthopedic trauma.

A study enrolling patients admitted to the Orthopedic Department who were subsequently transferred to the Intensive Care Unit (ICU) at Zhongda Hospital, Southeast University between January 2020 and December 2023, was conducted. Non-trauma patients and whose with an ICU length of stay (LOS) of less than 24 hours were excluded. Patients monitored by the early warning platform were assigned to the alert group (68 cases), while those receiving standard care formed the non-alert group (121 cases). The primary endpoint was hospital LOS. Multivariate linear regression was used to assess the independent association between the platform use and hospital LOS.

Among the 189 included patients, the median age was 72 years, 43.9% (83 cases) were female, 31.7% (60 cases) were emergency admissions, and 39.7% (75 cases) had spinal trauma. Baseline characteristics showed that, the proportion of patients with spinal trauma in the warning group (55.9%) was significantly higher than that in the non-alert group (30.6%), with a significant difference (χ²=10.612, P=0.001). Patients in the alert group had a significantly shorter median hospital LOS (P=0.006) and ICU LOS (P=0.017) compared to the non-alert group. Multivariate linear regression analysis confirmed that the application of the critical care early warning platform was independently associated with a reduction in hospital LOS (β=-5.91 days, SE=2.63, t=2.25, P=0.026).

The implementation of a real-time critical care early warning platform is associated with a significant reduction in hospital and ICU LOS among orthopedic trauma patients.

To translate and culturally adapt the self-report version of the healthy aging brain care monitor (HABC-MSR) into Chinese and test its reliability and validity in patients transferred from the intensive care unit (ICU), thereby, providing a simple and effective tool for the assessment of Post-Intensive Care Syndrome (PICS).

After obtaining authorization from the original scale developer, the HABC-MSR was translated and cross-culturally adapted according to the Brislin model. A convenience sample of 237 patients transferred from the intensive care unit of Peking Union Medical College Hospital to the general wards between April and December 2023 was enrolled. The reliability and validity of the Chinese version of the HABC-MSR were assessed.

The Chinese version of the HABC-MSR consists of 27 items across three dimensions: cognitive, physical, and psychological functioning. The scale-level content validity index (S-CVI/Ave) was 0.986, and the item-level content validity indices (I-CVI) ranged from 0.909 to 1.000. Confirmatory factor analysis demonstrated a good model fit. Criterion validity analysis revealed that the cognitive dimension showed no significant correlation with the Mini-Cog (r=-0.067, P=0.308). The physical function dimension was positively correlated with the Barthel Index (r=0.146, P<0.025), though the correlation was relatively weak. The psychological dimension showed a strong positive correlation with the Hospital Anxiety and Depression Scale (HADS) (r=0.598, P<0.001). The Cronbach's alpha coefficients of the subscales and the total scale ranged from 0.766 to 0.925, indicating good internal consistency.

The Chinese version of the HABC-MSR demonstrates good reliability and validity in patients transferred from ICU. It can serve as an effective assessment tool for PICS in Chinese clinical context.

To evaluate the effectiveness of healthcare failure mode and effect analysis (HFMEA) in improving the prevention and control of multidrug-resistant organism (MDRO) infections in critically ill neurosurgical patients.

The critically ill neurosurgical patients admitted to the ICU of Beijing Tiantan Hospital were divided into the control group (1726 cases from January to December 2023) and the experimental group (1812 cases from January to December 2024) according to their admission time. The control group received routine MDRO infection prevention and control measures; the experimental group received interventions based on an optimized process developed using the HFMEA model. Outcomes compared between the two groups included the incidence of MDRO infection, risk priority number (RPN) of failure modes, compliance with MDRO isolation precautions among nursing staff, and scores of infection prevention control culture questionnaire (LCOQ-IP).

Following the implemention of the HFMEA-based optimized process, the RPN values for 10 high-risk and medium-high-risk factors were significantly reduced (P<0.05). The MDRO infection rate in the experimental group was 7.45%, significantly lower than 11.88% in the control group (P<0.001). Compliance with isolation protocols-including the use of isolation signs, hand hygiene, isolation gowns, environmental disinfection, waste management, and terminal cleaning-was significantly higher in the experimental group compared to the control group (all P<0.05). Additionally, the LCOQ-IP score among ICU healthcare workers significantly increased after the intervention compared to before (P<0.05).

The HFMEA-based optimized MDRO management process for critically ill neurosurgical patients effectively reduces the incidence of MDRO infections, improves staff compliance with isolation precautions, and enhances the perceived culture of infection prevention and control among healthcare workers.

To conduct a bibliometric analysis of clinical research in the field of critical care medicine from 2020 to 2024, both internationally and in China, in order to identify current status, hotspots, and emerging trends.

A bibliometric methodology was employed. Clinical studies were analyzed using CiteSpace and VOSviewer software. The analysis was based on the Web of Science (WoS) core collection database, encompassing 8350 clinical studies published between January 1, 2020 and December 30, 2024. The analysis focused on international collaboration, patterns, author collaboration networks, keyword co-occurrence, and the characteristics of highly cited papers.

The study included 8350 clinical studies from 128 countries and regions. The United States led in publication output with 3563 articles. Analysis of the international collaboration network showed that the top three countries in publication volume were the United States, France, and the United Kingdom. The top three in terms of collaboration centrality were Northern Ireland, China, and Denmark. Keyword co-occurrence analysis identified major research hotspots including “Sepsis”, “Guidelines”, “Acute Respiratory Distress Syndrome”, “Children”， and “Mechanical ventilation”. Within the study period, 153 papers were identified as highly cited. The top 5 most cited papers (cited range: 761 to 5549) were published in the Lancet Respiratory Medicine and Intensive Care Medicine, with research themes concentrated on COVID-19 and Sepsis. The single highest internationally cited paper originated from China. The top 5 highly cited papers with Chinese researchers as participating (non-leading) authors had a citation range of 600 to 5549, while the top 5 papers with Chinese researchers as lead authors were cited 178 to 5549 times.

Countries including the United States, France, Australia, and Canada are the main driving forces in critical care medicine research and occupy central positions in the global scientific collaboration network. While a broad and extensive international cooperation network for China has not yet fully formed, its centrality within the existing network has significantly increased, indicating a growing influence.

To develop and validate an interpretable machine learning model for predicting the risk of prolonged mechanical ventilation (PMV) in adult patients requiring mechanical ventilation after cardiac valve surgery in ICU.

A retrospective cohort study of patients admitted to ICU after cardiac valve surgery who received mechanical ventilation at a tertiary hospital between July 1, 2023 and June 30, 2024, was conducted. Feature selection was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression. The dataset was randomly split into a training set and a validation set in a ratio of 7∶3. Multivariate logistic regression was used to adjust for potential confounders. Several ML models were evaluated and compared based on the area under the ROC curve (AUC), accuracy, sensitivity, specificity and F1 score. The optimal model was interpreted using SHapley Additive exPlanations (SHAP).

Among the 711 patients included, 143(20.1%) developed PMV. The logistic regression model demonstrated excellent discriminative ability, with an AUC for 0.903 (95%CI: 0.869-0.938) in the training set and 0.899 (95%CI: 0.854-0.943) in the validation set. SHAP analysis identified the following top predictors of PMV (in descending order of importance): lactate level ≥ 4 mmol/L at 6 hours post-operation, new-onset atrial fibrillation, cardiopulmonary bypass time ≥ 180 minutes, type of surgery, mean pulmonary artery pressure (MPAP) ≥ 35 mmHg, ejection fraction (EF) ≤ 0.35, partial pressure of oxygen < 90 mmHg, preoperative renal failure, re-do valve replacement, chest tube drainage ≥ 600 ml at 6 hours post-operation, and APACHE Ⅱ score≥ 21.

We developed an interpretable prediction model that performs well in identifying adult patients at high risk for PMV following cardiac valve surgery. The use of SHAP enhances the clinical utility of the model by providing insights into key risk factors, thereby aiding ICU clinicians in early risk stratification and potentially facilitating interventions to shorten the duration of mechanical ventilation.

Venoarterial extracorporeal membrane oxygenation (VA-ECMO) usually provides hemodynamic support for patients with cardiogenic shock (CS) and cardiac arrest, ensuring organ perfusion and leaving time for revascularization and reperfusion. Although ECMO technology is developing rapidly, it still faces many challenges in clinical practice. Through the analysis of key studies on VA-ECMO at home and abroad in recent years, this article summarizes the progress of VA-ECMO and its combination strategy, the key points of prevention and management of major complications, the standardized withdrawal process and prognosis evaluation, and explore the value of clinical application and development trend in future, aim to provide reference for optimizing individualized treatment decisions. In the future, we can focus on three major directions: etiological stratification research, complication technology research, and AI-driven individual decision-making. As a result, it will realize the shift from "generalized support" to "precise intervention", maximize the balance of risks and benefits and establish a more accurate ECMO management system.

The vasoactive-inotropic score (VIS) was developed to equivalently quantify the dosage of various vasoactive and inotropic agents, enabling a real-time, dynamic, direct, and objective assessment of disease severity, the intensity of pharmacological support, and hemodynamic status. In recent years, the VIS has been increasingly applied in clinical practice, particularly in the management of sepsis. This article reviews and synthesizes current evidence on the association between the VIS and the outcomes in sepsis, aiming to elucidate the prognostic value of the VIS. By consolidating this knowledge, this review seeks to assist clinicians in early identification of disease severity, prognostic stratification, and guidance of therapeutic strategies for sepsis patients.

The dysregulated inflammatory response is the primary pathophysiological mechanism underlying the onset of acute respiratory distress syndrome (ARDS) in septic patients. Distinct subpopulations of macrophages, namely monocyte-derived alveolar macrophages (Mo-AMs), tissue-resident alveolar macrophages (TRAMs), and pulmonary interstitial macrophages (PIMs), exhibit disparate functional profiles and exert varying impacts on the development of ARDS. Various therapeutic modalities have been proposed to target the three types of pulmonary macrophages mentioned above, encompassing interventions aimed at curtailing the generation of Mo-AMs, mitigating the hyperactivation of TRAM-mediated inflammation, and augmenting the anti-inflammatory capacities of PIMs. This review critically examines the functional paradigms of these pulmonary macrophage subtypes in the inflammatory pathogenesis of ARDS, thereby delineating avenues for further investigation and therapeutic exploration in ARDS management.

The nature of intensive care unit (ICU) often means family members cannot accompany their loved ones throughout the treatment process, which may lead to fear and misunderstanding. Therefore, enhancing public understanding of the ICU is crucial. "ICU Story: Simply Put" introduced in this article is a popular science book designed to bridge this gap. Featuring a cartoon spokesperson, "Peach Blossom Nine", and original illustrations, the book demystifies complex ICU-related knowledge through accessible language. By promoting a scientific understanding of medicine, it aims to bridge the gap between healthcare providers and the public, foster support for healthcare initiatives, and ultimately contribute to better outcomes for critically ill patients.

Paroxysmal sympathetic hyperactivity (PSH) is a rare complication of hemorrhagic stroke, thus, it could be easily misdiagnosed and overlooked. The clinical data of four PSH cases after hemorrhagic stroke were retrospectively analyzed. Patient 1, patient 3 and patient 4 suffered from severe aneurysmal subarachnoid hemorrhage. Patient 2 suffered from intracerebral hemorrhage with a large hematoma. PSH of patient 1 was related to a hematoma located in right frontal and insular lobe. PSH of patient 2 was related to a huge cyst which occupied right frontal, parietal, temporal and insular lobe, and right basal ganglia. In the case of patient 3, PSH was associated with cerebral herniation and the compressed midbrain. PSH of patient 4 was strongly related to the development of hydrocephalus. The PSH episodes of all patients responded well to sedation and analgesia treatment, or surgical treatment which could reduce the intracranial pressure, such as drainage or shunt. These cases indicate that both the structural and functional disconnection in insular lobe, diencephalon or brainstem within the central autonomic network could result in the development of PSH. Therefore, patients with lesions in these brain regions should be paid attention to the development of PSH.

Percutaneous dilatational tracheotomy (PDT), characterized by its minimal invasiveness and high safety profile, has become the mainstream technique for tracheotomy. To achieve adequate tracheal exposure, the procedure requires positioning the patient with shoulder elevation and neck hyperextension. However, this specific maneuver may increase intracranial pressure (ICP), which could potentially lead to adverse effects on long-term neurological outcomes. This case report presents the perioperative ICP monitoring data of a brain-death patient who underwent PDT approximately two weeks after acute brain injury. Our observations indicated that ICP remained persistently elevated by approximately 3-5 mmHg above the baseline from the establishment of the surgical position until the patient was returned to the initial supine position. Although this ICP elevation did not appear to directly influence the outcome in this particular case (given the patient's brain-death status), it nonetheless suggests a potential compromise to optimal ICP management. While early tracheostomy offers benefits such as enhanced patient comfort and reduced sedation requirements, balancing the timing of PDT in neurocritically ill patients against the risk of procedure-related ICP elevation and secondary neurological injury warrants further investigation through large-scale clinical studies.

Diquat is a non-selective, fast-acting herbicide belonging to the bipyridinium class of pesticides like paraquat. With the ban on paraquat, acute poisoning incidents involving diquat have been increasing yearly. However, the understanding of the injury mechanisms and clinical manifestations of acute diquat poisoning remains insufficient. This article reviewed the treatment of two cases of acute diquat poisoning leading to osmotic demyelination syndrome (ODS) in the Characteristic Medical Center of Chinese People's Armed Police Force, and summarized similar cases by searching Chinese and English databases including CNKI, Wanfang, and PubMed, using keywords "Diquat" and "Poisoning/Intoxication/Toxicity/Overdose." Among 12 patients, there were five males and seven females, with an average age of (32.6±16.3) years. Ingestion doses ranged from 50 to 120 ml, with coma being the most common clinical manifestation, reflecting significant impairment of consciousness. MRI findings typically showed low signal intensity on T1-weighted sequences, and high signal intensity on T2-weighted and FLAIR sequences in damaged areas including the pons, thalamus, and basal ganglia. Two patients died, while ten survived, with four survivors experiencing severe sequelae. This study suggests that in patients with acute diquat poisoning presenting with consciousness disorders such as coma, clinicians should be alert to the potential development of ODS and consider performing early MRI examination. While most of these patients have a good prognosis, some may develop severe sequelae, which warrants attention.

We reported a 67-year-old man with myelodysplastic syndrome who developed carbapenem-resistant Acinetobacter baumannii (CRAB)-induced severe pneumonia and sepsis after allogeneic hematopoietic stem cell transplantation. Post-transplantation, he presented with fever, hypoxemia and productive yellow sputum; chest CT showed bilateral diffuse interstitial inflammation with areas of consolidation. Repeated sputum cultures and blood metagenomic next-generation sequencing confirmed CRAB infection. Initial multidrug regimens - including polymyxin B, meropenem and eravacycline - failed to achieve clinical improvement. A Sulbactam-Durlobactam (SUL-DUR)-based combination regimen was then introduced, along with immunomodulatory therapy and airway management. Inflammatory markers (CRP and PCT) decreased rapidly, pathogens were eradicated, and the patient was successfully weaned from mechanical ventilation and transferred to the general ward. This case suggested that SUL-DUR-based therapy was both effective and well-tolerated for severe CRAB infections during immune reconstitution, and offered a promising therapeutic option for complex post-transplant infections.