Enhancing Nix-dependent mitophagy relieves AKI by restricting TREM-1-mediated hyperactivation of inflammasome in platelets.

Platelets are highly involved in inflammation and organ injury under pathological conditions. The mitophagy in platelets may restrict hyperactivation of the inflammasome and relieve acute kidney injury (AKI). Cecal ligation puncture (CLP)/LPS-induced AKI Triggering receptor expressed on myeloid cells (TREM-1)-knockout mice models were established. Additionally, septic patients with AKI were also included. TREM-1 expression in platelets and inflammasome activation were examined. Platelet transfer assays were performed to investigate the contribution of platelet TREM-1 to renal injury. Mitophagy was evaluated in the context of inflammation. BNIP3L/Nix knockout mice were used to examine the relationship between platelet mitophagy and inflammatory activation. The results showed that the level of TREM-1 was increased and the platelet inflammasome was hyperactivated in CLP mice and septic patients, and TREM-1 activated platelet inflammasomes. TREM-1 deletion significantly abrogated hyperactivation of the platelet inflammasome and dramatically reduced AKI, whereas ablation of the mitophagy receptor BNIP3L/Nix induced the accumulation of damaged mitochondria and hyperactivation of platelet inflammasomes in CLP mice. BNIP3L/Nix controlled platelet inflammasome activation, and an amplification loop of platelet inflammasome activation and dysfunctional mitochondria controlled sepsis-related AKI. Therefore, targeting TREM-1 and NLRP3/BNIP3L in platelets may represent a novel therapeutic strategy for treating septic AKI.

Factors influencing DVT formation in sepsis.

INTRODUCTION: Sepsis is a global public health burden. Deep vein thrombosis (DVT) is the third most common cause of death from cardiovascular disease after heart attacks and strokes. We designed this experiment to investigate the factors influencing DVT formation in patients with sepsis. METHODS: In this survey, 918 septic patients admitted to Peking Union Medical College Hospital, who underwent DVT screening were enrolled. The data were collected from June 8, 2013 to October 12, 2022. The differences between septic patients with and without DVT were studied from following aspects: basic information, comorbidities, inflammatory cytokines, albumin, source of infection, sequential organ failure assessment (SOFA) score, coagulation and prognosis. MAIN RESULTS: In this study, the prevalence of DVT in patients with sepsis was 0.23. Elderly patients with sepsis were prone to DVT (p value < 0.001). In terms of comorbidities, septic patients with hypertension and atrial fibrillation were prone to DVT (p value 0.045 and 0.048). Inflammatory cytokines, such as procalcitonin (PCT), C-reactive protein (CRP), interleukin (IL)-6, IL-8, IL-10, tumor necrosis factor (TNF)-α, had no significant correlation with DVT in patients with sepsis (p value 0.364, 0.882, 0.912, 0.789, 0.245, and 0.780). Levels of serum albumin correlated with DVT in patients with sepsis (p value 0.003). The SOFA total score had no relationship with DVT formation (p value 0.254). Coagulation and respiration function were negatively correlated with DVT (p value 0.018). Liver function was positively correlated with DVT (p value 0.020). Patients in the DVT group had longer duration of mechanical ventilation and longer intensive care unit (ICU) stays (p value < 0.001 and 0.006). There was no significant difference in survival in septic patients with and without DVT (p value 0.868). CONCLUSIONS: The SOFA total score had no relationship with DVT formation. The function of each organ had different effects on DVT formation. Better coagulation and respiration function, easier DVT formation. Poorer liver function, easier DVT formation. DVT was associated with longer duration of mechanical ventilation and longer ICU stays.

Blended teaching mode based on small private online course and case-based learning in analgesia and sedation education in China: a comparison with an offline mode.

BACKGROUND: Standardized training for pain, agitation-sedation, and delirium (PAD) management is urgently needed for Chinese intensivists' continuing education. Since 2020, because of the COVID-19 pandemic, the Chinese Analgesia and Sedation Education and Research (CASER) group has used an online blended teaching mode based on a small private online course (SPOC) and case-based learning (CBL). This study evaluated whether an online blended teaching mode has similar effects on PAD management training when an offline mode cannot be used. MATERIALS AND METHODS: Since 2020, the CASER group has provided offline training and online SPOC&CBL training three times each, targeting intensivists and ICU nurses in China. All participants were divided into an offline group and SPOC&CBL group. A final examination was offered in each training session to assess the students' mastery of professional knowledge. Teachers' and students' perceptions regarding the online SPOC&CBL mode were evaluated through questionnaires. RESULTS: Of all participants (n = 117), 106 completed all examinations and questionnaires. Most participants were aged 31-40 years (53, 50.0%), had an academic degree (60, 56.6%), and worked in a tertiary hospital (100, 94.34%). We assessed the learning effect on participants from two aspects: theory and clinical practice. There was no significant difference between the SPOC&CBL and offline groups in terms of theoretical, case analysis, and total scores (p > 0.05). In terms of the participants' perceptions regarding the SPOC&CBL mode, 91.5% considered the online mode to be a useful and accessible alternative to improve knowledge and skills. A total of 95.7% of the participants believed that they could interact well with group members, and 87.2% believed that they had a good degree of participation. Of these participants, 76.6% believed that they had received valuable learning resources. All instructors believed that the SPOC&CBL mode was more flexible than the offline mode in terms of teaching time and location, and they were all willing to carry out training with the SPOC&CBL mode. CONCLUSION: Compared to the offline mode, the SPOC&CBL mode can also enhance participants' knowledge and skills and meets their expectations. Therefore, an online mode can be considered a potential method in PAD management education in China.

Tissue oxygen saturation is predictive of lactate clearance in patients with circulatory shock.

BACKGROUND: Tissue oxygen saturation (StO2) decrease could appear earlier than lactate alteration. However, the correlation between StO2 and lactate clearance was unknown. METHODS: This was a prospective observational study. All consecutive patients with circulatory shock and lactate over 3 mmol/L were included. Based on the rule of nines, a BSA (body surface area) weighted StO2 was calculated from four sites of StO2 (masseter, deltoid, thenar and knee). The formulation was as follows: masseter StO2 × 9% + (deltoid StO2 + thenar StO2) × (18% + 27%)/ 2 + knee StO2 × 46%. Vital signs, blood lactate, arterial and central venous blood gas were measured simultaneously within 48 h of ICU admission. The predictive value of BSA-weighted StO2 on 6-hour lactate clearance > 10% since StO2 initially monitored was assessed. RESULTS: A total of 34 patients were included, of whom 19 (55.9%) had a lactate clearance higher than 10%. The mean SOFA score was lower in cLac ≥ 10% group compared with cLac < 10% group (11 ± 3 vs. 15 ± 4, p = 0.007). Other baseline characteristics were comparable between groups. Compared to non-clearance group, StO2 in deltoid, thenar and knee were significantly higher in clearance group. The area under the receiver operating curves (AUROC) of BSA-weighted StO2 for prediction of lactate clearance (0.92, 95% CI [Confidence Interval] 0.82-1.00) was significantly higher than StO2 of masseter (0.65, 95% CI 0.45-0.84; p < 0.01), deltoid (0.77, 95% CI 0.60-0.94; p = 0.04), thenar (0.72, 95% CI 0.55-0.90; p = 0.01), and similar to knee (0.87, 0.73-1.00; p = 0.40), mean StO2 (0.85, 0.73-0.98; p = 0.09). Additionally, BSA-weighted StO2 model had continuous net reclassification improvement (NRI) over the knee StO2 and mean StO2 model (continuous NRI 48.1% and 90.2%, respectively). The AUROC of BSA-weighted StO2 was 0.91(95% CI 0.75-1.0) adjusted by mean arterial pressure and norepinephrine dose. CONCLUSIONS: Our results suggested that BSA-weighted StO2 was a strong predictor of 6-hour lactate clearance in patients with shock.

Measuring the worldwide spread of COVID-19 using a comprehensive modeling method.

BACKGROUND: With the global spread of COVID-19, detecting high-risk countries/regions timely and dynamically is essential; therefore, we sought to develop automatic, quantitative and scalable analysis methods to observe and estimate COVID-19 spread worldwide and further generate reliable and timely decision-making support for public health management using a comprehensive modeling method based on multiple mathematical models. METHODS: We collected global COVID-19 epidemic data reported from January 23 to September 30, 2020, to observe and estimate its possible spread trends. Countries were divided into three outbreak levels: high, middle, and low. Trends analysis was performed by calculating the growth rate, and then country grouping was implemented using group-based trajectory modeling on the three levels. Individual countries from each group were also chosen to further disclose the outbreak situations using two predicting models: the logistic growth model and the SEIR model. RESULTS: All 187 observed countries' trajectory subgroups were identified using two grouping strategies: with and without population consideration. By measuring epidemic trends and predicting the epidemic size and peak of individual countries, our study found that the logistic growth model generally estimated a smaller epidemic size than the SEIR model. According to SEIR modeling, confirmed cases in each country would take an average of 9-12 months to reach the outbreak peak from the day the first case occurred. Additionally, the average number of cases at the peak time will reach approximately 10-20% of the countries' populations, and the countries with high trends and a high predicted size must pay special attention and implement public health interventions in a timely manner. CONCLUSIONS: We demonstrated comprehensive observations and predictions of the COVID-19 outbreak in 187 countries using a comprehensive modeling method. The methods proposed in this study can measure COVID-19 development from multiple perspectives and are generalizable to other epidemic diseases. Furthermore, the methods also provide reliable and timely decision-making support for public health management.

Importance of clinical parameters for cultivation of critical care thinking by online teaching of critical care medicine.

BACKGROUND: The teaching of critical care medicine is a very important task, especially during the COVID-19 pandemic. The understanding of critical care parameters is the foundation and core, which is conducive to the formation of clinical thinking. This study is to evaluate the training effect of teaching of critical care parameters based on an online platform, and explore the teaching methods of critical care medicine that can help to cultivate trainees' clinical thinking and practical ability. METHODS: Questionnaires were released before and after the training through the official new media platform "Yisheng" application (APP) of China Medical Tribune involving 1109 participants. The trainees who filled in the questionnaire in APP and received training were randomly selected as the investigated population. Statistical description and analysis were carried out using SPSS 20.0 and Excel 2020. RESULTS: The trainees were mainly attending physicians in tertiary hospitals and above. Among all critical care parameters, trainees paid more attention to critical hemodynamics, respiratory mechanics, severity of illness scoring systems, critical ultrasound, and critical hemofiltration. The degree of satisfaction with the courses was high, especially the course of critical hemodynamics was scored the highest. The trainees believed that the course contents were of great help to clinical work. However, no significant difference was found in the trainees' understanding or cognition of the connotation of the parameters before and after the training. CONCLUSION: Teaching of critical care parameters based on an online platform is conducive to improving and consolidating the clinical care ability of trainees. However, it is still necessary to strengthen the cultivation of clinical thinking in critical care. In the future, the integration of theory with practice must be strengthened in clinical practice, ultimately achieving the homogeneous diagnosis and treatment of patients with critical illness.

Significance of platelets in the early warning of new-onset AKI in the ICU by using supervise learning: a retrospective analysis.

OBJECTIVE: To explore a machine learning model for the early prediction of acute kidney injury (AKI) and to screen the related factors affecting new-onset AKI in the ICU. METHODS: A retrospective analysis was performed used the MIMIC-III data source. New onset of AKI defined based on the serum creatinine changed. We included 19 variables for AKI assessment using four machine learning models: support vector machines, logistic regression, and random forest. and XGBoost, using accuracy, specificity, precision, recall, F1 score, and AUROC (area under the ROC curve) to evaluate model performance. The four models predicted new-onset AKI 3-6-9-12 h ahead. The SHapley Additive exPlanation (SHAP) value is used to evaluate the feature importance of the model. RESULTS: We finally respectively extracted 1130 AKI patients and non-AKI patients from the MIMIC-III database. With the extension of the early warning time, the prediction performance of each model showed a downward trend, but the relative performance was consistent. The prediction performance comparison of the four models showed that the XGBoost model performed the best in all evaluation indicators in all the time point at new-onset AKI 3-6-9-12 h ahead (accuracy 0.809 vs 0.78 vs 0.744 vs 0.741, specificity 0.856 vs 0.826 vs 0.797 vs 0.787, precision 0.842 vs 0.81 vs 0.775 vs 0.766, recall 0.759 vs 0.734 vs 0.692 vs 0.694, Fl score 0.799 vs 0.769 vs 0.731 vs 0.729, AUROC 0.892 vs 0.857 vs 0.827 vs 0.818). In the prediction of AKI 6, 9 and 12 h ahead, the importance of creatinine, platelets, and height was the most important based on SHapley. CONCLUSIONS: The machine learning model described in this study can predict AKI 3-6-9-12 h before the new-onset of AKI in ICU. In particular, platelet plays an important role.

The new-onset of AKI in ICU is a common and important problem, which early be identified the risk of AKI can improve patients' outcomes.We explored MIMIC-III and determined the exact time point of occurrence of AKI as the basis for the new-onset of AKI in ICU.XGBoost model performed the best prediction in all the time point at new-onset AKI 3­6­9­12 h ahead.For patients with the new-onset of AKI in ICU, platelets become an important factor associated with AKI.

Association between hospital and ICU structural factors and patient outcomes in China: a secondary analysis of the National Clinical Improvement System Data in 2019.

BACKGROUND: Hospital and ICU structural factors are key factors affecting the quality of care as well as ICU patient outcomes. However, the data from China are scarce. This study was designed to investigate how differences in patient outcomes are associated with differences in hospital and ICU structure variables in China throughout 2019. METHODS: This was a multicenter observational study. Data from a total of 2820 hospitals were collected using the National Clinical Improvement System Data that reports ICU information in China. Data collection consisted of a) information on the hospital and ICU structural factors, including the hospital type, number of beds, staffing, among others, and b) ICU patient outcomes, including the mortality rate as well as the incidence of ventilator-associated pneumonia (VAP), catheter-related bloodstream infections (CRBSIs), and catheter-associated urinary tract infections (CAUTIs). Generalized linear mixed models were used to analyse the association between hospital and ICU structural factors and patient outcomes. RESULTS: The median ICU patient mortality was 8.02% (3.78%, 14.35%), and the incidences of VAP, CRBSI, and CAUTI were 5.58 (1.55, 11.67) per 1000 ventilator days, 0.63 (0, 2.01) per 1000 catheter days, and 1.42 (0.37, 3.40) per 1000 catheter days, respectively. Mortality was significantly lower in public hospitals (ß = - 0.018 (- 0.031, - 0.005), p = 0.006), hospitals with an ICU-to-hospital bed percentage of more than 2% (ß = - 0.027 (- 0.034, -0.019), p < 0.001) and higher in hospitals with a bed-to-nurse ratio of more than 0.5:1 (ß = 0.009 (0.001, 0.017), p = 0.027). The incidence of VAP was lower in public hospitals (ß = - 0.036 (- 0.054, - 0.018), p < 0.001). The incidence of CRBSIs was lower in public hospitals (ß = - 0.008 (- 0.014, - 0.002), p = 0.011) and higher in secondary hospitals (ß = 0.005 (0.001, 0.009), p = 0.010), while the incidence of CAUTIs was higher in secondary hospitals (ß = 0.010 (0.002, 0.018), p = 0.015). CONCLUSION: This study highlights the association between specific ICU structural factors and patient outcomes. Modifying structural factors is a potential opportunity that could improve patient outcomes in ICUs.

Effect of ICU quality control indicators on VAP incidence rate and mortality: a retrospective study of 1267 hospitals in China.

PURPOSE: To investigate the effects of ICU quality control indicators on the VAP incidence rate and mortality in China throughout 2019. METHODS: This was a retrospective study. A total of 1267 ICUs from 30 provinces in mainland China were included. Data were collected using the National Clinical Improvement System Data that report ICU information. Ten related quality control indicators were analyzed, including 5 structural factors (patient-to-bed ratio, physician-to-bed ratio, nurse-to-bed ratio, patient-to-physician ratio, and patient-to-nurse ratio), 3 process factors (unplanned endotracheal extubation rate, reintubation rate within 48 h, and microbiology detection rate before antibiotic use), and 2 outcome factors (VAP incidence rate and mortality). The information on the most common infectious pathogens and the most commonly used antibiotics in ICU was also collected. The Poisson regression model was used to identify the impact of factors on the incidence rate and mortality of VAP. RESULTS: The incidence rate of VAP in these hospitals in 2019 was 5.03 (2.38, 10.25) per 1000 ventilator days, and the mortality of VAP was 11.11 (0.32, 26.00) %. The most common causative pathogen was Acinetobacter baumannii (in 39.98% of hospitals), followed by Klebsiella pneumoniae (38.26%), Pseudomonas aeruginosa, and Escherichia coli. In 26.90% of hospitals, third-generation cephalosporin was the most used antibiotic, followed by carbapenem (24.22%), penicillin and beta-lactamase inhibitor combination (20.09%), cephalosporin with beta-lactamase inhibitor (17.93%). All the structural factors were significantly associated with VAP incidence rate, but not with the mortality, although the trend was inconsistent. Process factors including unplanned endotracheal extubation rate, reintubation rate in 48 h, and microbiology detection rate before antibiotic use were associated with higher VAP mortality, while unplanned endotracheal extubation rate and reintubation rate in 48 h were associated with higher VAP mortality. Furthermore, K. pneumoniae as the most common pathogen was associated with higher VAP mortality, and carbapenems as the most used antibiotics were associated with lower VAP mortality. CONCLUSION: This study highlights the association between the ICU quality control (QC) factors and VAP incidence rate and mortality. The process factors rather than the structural factors need to be further improved for the QC of VAP in the ICU.

The relationship between ventilatory ratio (VR) and 28-day hospital mortality by restricted cubic splines (RCS) in 14,328 mechanically ventilated ICU patients.

BACKGROUND: Previous studies found that high levels of ventilatory ratio (VR) were associated with a poor prognosis due to worse ventilatory efficiency in acute respiratory distress syndrome patients. However, relatively few large studies have assessed the association between VR and intensive care unit (ICU) mortality in the general adult ventilated population. METHODS: The present study is a retrospective cohort study. Patients mechanically ventilated for more than 12 h were included. VR was calculated based on a previously reported formula. Restricted cubic spline models were used to fit the relationship between VR and mortality risks. RESULTS: A total of 14,328 mechanically ventilated ICU patients were included in the study, of which 1311 died within 28 days. The results of the study are as follows: (1) In the general adult ventilated population, VR was positively associated with 28-day mortality when VR ≥ 1.3 (increase of 0.1 per VR; HR 1.05, p < 0.001). The same tendency was also observed in the populations of severe hypoxemia with a PaO2/FiO2 (P/F) ratio < 200 mmHg. (2) However, in the population with a P/F ratio ≥ 200, a J-shaped dose-response association between VR and the risk of mortality was observed, with the risk of death positively associated with VR when VR ≥ 0.9 (10% increase in HR for every 0.1 increase in VR, p = 0.000) but negatively associated with VR when VR < 0.9 (10% decrease in HR for every 0.1 increase in VR, p = 0.034). In the population of P/F ratio ≥ 200 with VR less than 0.9, compared to the survival group, the nonsurvival group had a lower level PCO2 (33 mmHg [29.1, 37.9] vs. 34.4 mmHg [30.6, 38.5]), rather than a significant level of measured minute ventilation or P/F ratio. CONCLUSIONS: VR was positively associated with the risk of death in the general ICU population; however, VR was inversely associated with 28-day mortality in the population with a P/F ratio ≥ 200 and low VR . Further research should investigate this relationship, and VR should be interpreted with caution in clinical practice.

TREM-1 promoted apoptosis and inhibited autophagy in LPS-treated HK-2 cells through the NF-κB pathway.

Triggering receptor expressed by myeloid cells (TREM-1) is an amplifier of inflammatory responses triggered by bacterial or fungal infection. Soluble TREM-1 (sTREM-1) expression was found to be upregulated in sepsis-associated acute kidney injury (SA-AKI) and predicted to be a potential biomarker. However, the mechanism remains unclear. The human kidney-2 (HK-2) cell line was treated with lipopolysaccharide (LPS) and used to examine the potential roles of TREM-1 in apoptosis and autophagy. A cell viability assay was employed to assess the number of viable cells and as a measure of the proliferative index. The concentrations of sTREM-1, interleukin (IL)-1ß, tumor necrosis factor-α (TNFα) and IL-6 in cell-free culture supernatants were measured by enzyme-linked immunosorbent assay (ELISA). Western blot analysis was performed to analyze apoptosis, autophagy and the relevant signaling pathways. The results suggested that TREM-1 overexpression after LPS treatment decreased proliferation and increased apoptosis. The concentrations of sTREM-1, IL-1ß, TNFα and IL-6 in cell-free culture supernatants were increased in the TREM-1 overexpression group after LPS treatment. Expression of the antiapoptotic gene Bcl-2 was downregulated in the TREM-1 overexpression group, while that of the proapoptotic genes Bax, cleaved caspase-3 and cleaved caspase-9 was upregulated. Overexpression of TREM-1 downregulated expression of the autophagy genes Beclin-1, Atg-5 and LC3b and increased the gene expression of p62, which inhibits autophagy. Conversely, treatment with TREM-1-specific shRNA had the opposite effects. The nuclear factor-κB (NF-κB) signaling pathway (P-p65/p65 and P-IκBα/IκBα) in LPS-induced HK-2 cells was regulated by TREM-1. In summary, TREM-1 promoted apoptosis and inhibited autophagy in HK-2 cells in the context of LPS exposure potentially through the NF-κB pathway.

Negative central venous to arterial lactate gradient in patients receiving vasopressors is associated with higher ICU 30-day mortality: a retrospective cohort study.

BACKGROUND: Serum lactate has long been used to evaluate hypoxia and predict prognosis in critically ill patients, however, discrepancy in lactate measurements between different sites have not been recognized as a useful tool for monitoring hypoxia and evaluating outcome. METHODS: Data were obtained from the clinical information system of the intensive care unit (ICU) in a tertiary academic hospital for 1582 ICU patients with vasoactive drug requirement and valid paired blood gas. The mortality rates were compared between patients with sustained negative venous to arterial lactate gradient (VALac) and the others using the Cox proportional hazard model. Predictive factors associated with negative VALac were searched. RESULTS: A sustained negative VALac was significantly associated with higher 30 day ICU mortality [Adjusted hazard ratio (HR) = 2.31, 95% confidence interval (CI), 1.07-4.99; p = 0.032. Propensity score- weighted HR: 2.57; 95% CI, 1.17-5.64; p = 0.010]. Arterial lactate in the first blood gas pair, 24-h arterial lactate clearance, use of epinephrine, mean positive end-expiratory pressure level, and extracorporeal membrane oxygenation initiation showed statistically significant association with sustained negative VALac during the first 24 h. CONCLUSION: The sustained negative VALac in the early stage of treatment may suggest additional information about tissue hypoxia than arterial lactate alone. Critical care physicians should pay more attention to the lactate discrepancy between different sites in their clinical practice.

Stepwise lactate kinetics in critically ill patients: prognostic, influencing factors, and clinical phenotype.

BACKGROUND: To investigate the optimal target e of lactate kinetics at different time during the resuscitation, the factors that influence whether the kinetics achieve the goals, and the clinical implications of different clinical phenotypes. METHODS: Patients with hyperlactatemia between May 1, 2013 and December 31, 2018 were retrospectively analyzed. Demographic data, basic organ function, hemodynamic parameters at ICU admission (T0) and at 6 h, 12 h, 24 h, 48 h, and 72 h, arterial blood lactate and blood glucose levels, cumulative clinical treatment conditions at different time points and final patient outcomes were collected. RESULTS: A total of 3298 patients were enrolled, and the mortality rate was 12.2%. The cutoff values of lactate kinetics for prognosis at 6 h, 12 h, 24 h, 48 h, and 72 h were 21%, 40%, 57%, 66%, and 72%. The APACHE II score, SOFA score, heart rate (HR), and blood glucose were risk factors that correlated with whether the lactate kinetics attained the target goal. Based on the pattens of the lactate kinetics, eight clinical phenotypes were proposed. The odds ratios of death for clinical phenotypes VIII, IV, and II were 4.39, 4.2, and 5.27-fold of those of clinical phenotype I, respectively. CONCLUSION: Stepwise recovery of lactate kinetics is an important resuscitation target for patients with hyperlactatemia. The APACHE II score, SOFA score, HR, and blood glucose were independent risk factors that influenced achievement of lactate kinetic targets. The cinical phenotypes of stepwise lactate kinetics are closely related to the prognosis.

A Clinical Prediction Model to Predict Heparin Treatment Outcomes and Provide Dosage Recommendations: Development and Validation Study.

BACKGROUND: Unfractionated heparin is widely used in the intensive care unit as an anticoagulant. However, weight-based heparin dosing has been shown to be suboptimal and may place patients at unnecessary risk during their intensive care unit stay. OBJECTIVE: In this study, we intended to develop and validate a machine learning-based model to predict heparin treatment outcomes and to provide dosage recommendations to clinicians. METHODS: A shallow neural network model was adopted in a retrospective cohort of patients from the Multiparameter Intelligent Monitoring in Intensive Care III (MIMIC III) database and patients admitted to the Peking Union Medical College Hospital (PUMCH). We modeled the subtherapeutic, normal, and supratherapeutic activated partial thromboplastin time (aPTT) as the outcomes of heparin treatment and used a group of clinical features for modeling. Our model classifies patients into 3 different therapeutic states. We tested the prediction ability of our model and evaluated its performance by using accuracy, the kappa coefficient, precision, recall, and the F1 score. Furthermore, a dosage recommendation module was designed and evaluated for clinical decision support. RESULTS: A total of 3607 patients selected from MIMIC III and 1549 patients admitted to the PUMCH who met our criteria were included in this study. The shallow neural network model showed results of F1 scores 0.887 (MIMIC III) and 0.925 (PUMCH). When compared with the actual dosage prescribed, our model recommended increasing the dosage for 72.2% (MIMIC III, 1240/1718) and 64.7% (PUMCH, 281/434) of the subtherapeutic patients and decreasing the dosage for 80.9% (MIMIC III, 504/623) and 76.7% (PUMCH, 277/361) of the supratherapeutic patients, suggesting that the recommendations can contribute to clinical improvements and that they may effectively reduce the time to optimal dosage in the clinical setting. CONCLUSIONS: The evaluation of our model for predicting heparin treatment outcomes demonstrated that the developed model is potentially applicable for reducing the misdosage of heparin and for providing appropriate decision recommendations to clinicians.

Early warning of citric acid overdose and timely adjustment of regional citrate anticoagulation based on machine learning methods.

BACKGROUND: Regional citrate anticoagulation (RCA) is an important local anticoagulation method during bedside continuous renal replacement therapy. To improve patient safety and achieve computer assisted dose monitoring and control, we took intensive care units patients into cohort and aiming at developing a data-driven machine learning model to give early warning of citric acid overdose and provide adjustment suggestions on citrate pumping rate and 10% calcium gluconate input rate for RCA treatment. METHODS: Patient age, gender, pumped citric acid dose value, 5% NaHCO3 solvent, replacement fluid solvent, body temperature value, and replacement fluid PH value as clinical features, models attempted to classify patients who received regional citrate anticoagulation into correct outcome category. Four models, Adaboost, XGBoost, support vector machine (SVM) and shallow neural network, were compared on the performance of predicting outcomes. Prediction results were evaluated using accuracy, precision, recall and F1-score. RESULTS: For classifying patients at the early stages of citric acid treatment, the accuracy of neutral networks model is higher than Adaboost, XGBoost and SVM, the F1-score of shallow neutral networks (90.77%) is overall outperformed than other models (88.40%, 82.17% and 88.96% for Adaboost, XGBoost and SVM). Extended experiment and validation were further conducted using the MIMIC-III database, the F1-scores for shallow neutral networks, Adaboost, XGBoost and SVM are 80.00%, 80.46%, 80.37% and 78.90%, the AUCs are 0.8638, 0.8086, 0.8466 and 0.7919 respectively. CONCLUSION: The results of this study demonstrated the feasibility and performance of machine learning methods for monitoring and adjusting local regional citrate anticoagulation, and further provide decision-making recommendations to clinicians point-of-care.

Selection strategy for sedation depth in critically ill patients on mechanical ventilation.

BACKGROUND: Analgesia and sedation therapy are commonly used for critically ill patients, especially mechanically ventilated patients. From the initial nonsedation programs to deep sedation and then to on-demand sedation, the understanding of sedation therapy continues to deepen. However, according to different patient's condition, understanding the individual patient's depth of sedation needs remains unclear. METHODS: The public open source critical illness database Medical Information Mart for Intensive Care III was used in this study. Latent profile analysis was used as a clustering method to classify mechanically ventilated patients based on 36 variables. Principal component analysis dimensionality reduction was used to select the most influential variables. The ROC curve was used to evaluate the classification accuracy of the model. RESULTS: Based on 36 characteristic variables, we divided patients undergoing mechanical ventilation and sedation and analgesia into two categories with different mortality rates, then further reduced the dimensionality of the data and obtained the 9 variables that had the greatest impact on classification, most of which were ventilator parameters. According to the Richmond-ASS scores, the two phenotypes of patients had different degrees of sedation and analgesia, and the corresponding ventilator parameters were also significantly different. We divided the validation cohort into three different levels of sedation, revealing that patients with high ventilator conditions needed a deeper level of sedation, while patients with low ventilator conditions required reduction in the depth of sedation as soon as possible to promote recovery and avoid reinjury. CONCLUSION: Through latent profile analysis and dimensionality reduction, we divided patients treated with mechanical ventilation and sedation and analgesia into two categories with different mortalities and obtained 9 variables that had the greatest impact on classification, which revealed that the depth of sedation was limited by the condition of the respiratory system.

IP-10 and MCP-1 as biomarkers associated with disease severity of COVID-19.

BACKGROUND: COVID-19 is a viral respiratory disease caused by the severe acute respiratory syndrome-Coronavirus type 2 (SARS-CoV-2). Patients with this disease may be more prone to venous or arterial thrombosis because of the activation of many factors involved in it, including inflammation, platelet activation and endothelial dysfunction. Interferon gamma inducible protein-10 (IP-10), monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein 1-alpha (MIP1α) are cytokines related to thrombosis. Therefore, this study focused on these three indicators in COVID-19, with the hope to find biomarkers that are associated with patients' outcome. METHODS: This is a retrospective single-center study involving 74 severe and critically ill COVID-19 patients recruited from the ICU department of the Tongji Hospital in Wuhan, China. The patients were divided into two groups: severe patients and critically ill patients. The serum IP-10, MCP-1 and MIP1α level in both groups was detected using the enzyme-linked immunosorbent assay (ELISA) kit. The clinical symptoms, laboratory test results, and the outcome of COVID-19 patients were retrospectively analyzed. RESULTS: The serum IP-10 and MCP-1 level in critically ill patients was significantly higher than that in severe patients (P < 0.001). However, no statistical difference in MIP1α between the two groups was found. The analysis of dynamic changes showed that these indicators remarkably increased in patients with poor prognosis. Since the selected patients were severe or critically ill, no significant difference was observed between survival and death. CONCLUSIONS: IP-10 and MCP-1 are biomarkers associated with the severity of COVID-19 disease and can be related to the risk of death in COVID-19 patients.

Five novel clinical phenotypes for critically ill patients with mechanical ventilation in intensive care units: a retrospective and multi database study.

BACKGROUND: Although protective mechanical ventilation (MV) has been used in a variety of applications, lung injury may occur in both patients with and without acute respiratory distress syndrome (ARDS). The purpose of this study is to use machine learning to identify clinical phenotypes for critically ill patients with MV in intensive care units (ICUs). METHODS: A retrospective cohort study was conducted with 5013 patients who had undergone MV and treatment in the Department of Critical Care Medicine, Peking Union Medical College Hospital. Statistical and machine learning methods were used. All the data used in this study, including demographics, vital signs, circulation parameters and mechanical ventilator parameters, etc., were automatically extracted from the electronic health record (EHR) system. An external database, Medical Information Mart for Intensive Care III (MIMIC III), was used for validation. RESULTS: Phenotypes were derived from a total of 4009 patients who underwent MV using a latent profile analysis of 22 variables. The associations between the phenotypes and disease severity and clinical outcomes were assessed. Another 1004 patients in the database were enrolled for validation. Of the five derived phenotypes, phenotype I was the most common subgroup (n = 2174; 54.2%) and was mostly composed of the postoperative population. Phenotype II (n = 480; 12.0%) led to the most severe conditions. Phenotype III (n = 241; 6.01%) was associated with high positive end-expiratory pressure (PEEP) and low mean airway pressure. Phenotype IV (n = 368; 9.18%) was associated with high driving pressure, and younger patients comprised a large proportion of the phenotype V group (n = 746; 18.6%). In addition, we found that the mortality rate of Phenotype IV was significantly higher than that of the other phenotypes. In this subgroup, the number of patients in the sequential organ failure assessment (SOFA) score segment (9,22] was 198, the number of deaths was 88, and the mortality rate was higher than 44%. However, the cumulative 28-day mortality of Phenotypes IV and II, which were 101 of 368 (27.4%) and 87 of 480 (18.1%) unique patients, respectively, was significantly higher than those of the other phenotypes. There were consistent phenotype distributions and differences in biomarker patterns by phenotype in the validation cohort, and external verification with MIMIC III further generated supportive results. CONCLUSIONS: Five clinical phenotypes were correlated with different disease severities and clinical outcomes, which suggested that these phenotypes may help in understanding heterogeneity in MV treatment effects.

Effects of a national quality improvement program on ICUs in China: a controlled pre-post cohort study in 586 hospitals.

INTRODUCTION: Patient safety and critical care quality remain a challenging issue in the ICU. However, the effects of the national quality improvement (QI) program remain unknown in China. METHODS: A national ICU QI program was implemented in a controlled cohort of 586 hospitals from 2016 to 2018. The effects of the QI program on critical care quality were comprehensively investigated. MAIN RESULTS: A total of 81,461,554 patients were enrolled in 586 hospitals, and 1,587,724 patients were admitted to the ICU over 3 years. In 2018, there was a significantly higher number of ICU beds (2016 vs. 2018: 10668 vs. 13,661, P = 0.0132) but a lower doctor-to-bed ratio (2016 vs. 2018: 0.64 (0.50, 0.83) vs. 0.60 (0.45, 0.75), P = 0.0016) and nurse-to-bed ratio (2016 vs. 2018: 2.00 (1.64, 2.50) vs. 2.00 (1.50, 2.40), P = 0.031) than in 2016. Continuous and significant improvements in the ventilator-associated pneumonia (VAP) incidence rate, microbiology detection rate before antibiotic use and deep vein thrombosis (DVT) prophylaxis rate were associated with the implementation of the QI program (VAP incidence rate (per 1000 ventilator-days), 2016 vs. 2017 vs. 2018: 11.06 (4.23, 22.70) vs. 10.20 (4.25, 23.94) vs. 8.05 (3.13, 17.37), P = 0.0002; microbiology detection rate before antibiotic use (%), 2016 vs. 2017 vs. 2018: 83.91 (49.75, 97.87) vs. 84.14 (60.46, 97.24) vs. 90.00 (69.62, 100), P < 0.0001; DVT prophylaxis rate, 2016 vs. 2017 vs. 2018: 74.19 (33.47, 96.16) vs. 71.70 (38.05, 96.28) vs. 83.27 (47.36, 97.77), P = 0.0093). Moreover, the 6-h SSC bundle compliance rates in 2018 were significantly higher than those in 2016 (6-h SSC bundle compliance rate, 2016 vs. 2018: 64.93 (33.55, 93.06) vs. 76.19 (46.88, 96.67)). A significant change trend was not found in the ICU mortality rate from 2016 to 2018 (ICU mortality rate (%), 2016 vs. 2017 vs. 2018: 8.49 (4.42, 14.82) vs. 8.95 (4.89, 15.70) vs. 9.05 (5.12, 15.80), P = 0.1075). CONCLUSIONS: The relationship between medical human resources and ICU overexpansion was mismatched during the past 3 years. The implementation of a national QI program improved ICU performance but did not reduce ICU mortality.

Cross-sectional study for the clinical application of extracorporeal membrane oxygenation in Mainland China, 2018.

BACKGROUND: To investigate the epidemiology and in-hospital mortality of veno-venous (VV) and veno-arterial (VA) extracorporeal membrane oxygenation (ECMO) in Mainland China throughout 2018. METHODS: Patients supported by ECMO from 1700 tertiary hospitals in 31 provinces from January 1 to December 31, 2018, were selected from the National Clinical Improvement System database. RESULTS: The 1700 included hospitals had 2073 cases of ECMO in 2018, including 714 VV and 1359 VA ECMOs. The average patient age was 50 years (IQR 31-63), and 1346 were male. The average hospital stay was 17 days (IQR 7-30), and the average costs per case was $36,334 (IQR 22,547-56,714). The three provinces with the highest number of ECMO cases were Guangdong, Beijing, and Zhejiang; the southeast coastal areas and regions with higher GDP levels had more cases. Overall in-hospital mortality was 29.6%. Mortality was higher among patients who were male, over 70 years old, living in underdeveloped areas, and who were treated during the summer. Mortality in provinces with more ECMO cases was relatively low. The co-existence of congenital malformations, blood system abnormalities, or nervous system abnormalities increased in-hospital mortality. CONCLUSIONS: Mortality and medical expenses of ECMO among patients in China were relatively low, but large regional and seasonal differences were present. Risk factors for higher in-hospital mortality were older age, male sex, in underdeveloped areas, and treatment during the summer. Additionally, congenital malformations and blood system and nervous system abnormalities were associated with in-hospital mortality.