Background and Objectives: Sepsis is one of the common causes of death in intensive care units. A reliable prognostic model based on patients' data acquired at the intensive care unit (ICU) would enable clinicians to make treatment decisions to improve clinical outcomes for septic patients. This study aims to develop a machine-learning framework for building such prognostic tools by exploring the class-imbalanced longitudinal data of a group of septic patients.Methods: A feature-represented input dataset is devised in the form of concatenated triples to increase the data size relative to the dimension of the feature space. Each concatenated triplet consists of a patient's static data, the k-day consecutively collected longitudinal data, and the clinical outcome (k=2,3,4,5). The structured input data are then used to train classifiers in combination with appropriate feature engineering techniques. The trained classifiers are tested on a new set of septic patients to ensure their clinical efficacy. We implement the modeling approach using five classifiers: K nearest neighbors, Logistic Regression, Support Vector Machine, Random Forest (RF), and Extreme Gradient Boosting (XGBoost) coupled with a set of feature engineering techniques. AUROC and a new metric, $\gamma$, made up of the F1 score on the external validation set, are used to assess the efficacy of the models.Results: Five prognostic models are built on the feature-represented input dataset accounting for 10 selected dynamic features from the patient medical data. Our research shows that the XGBoost (AUROC=0.777, F1 score=0.694) and RF (AUROC=0.769, F1 score=0.647) model combined with the ensemble under-sampling strategy outperform all other models in the external validation or testing. For example, the improvement in AUROC and overfitting are (6.66\%, 54.96\%) and (0.52\%, 77.72\%) for the RF and XGBoost model with the sampling strategy compared to the same models without using the sampling strategy, respectively. This indicates that the machine-learning framework can greatly improve the accuracy and generalizability of standard classifiers. Conclusion:A new modeling framework is devised to develop prognostic tools for treatment outcomes of septic patients using small, class-imbalanced, and high-dimensional datasets. It enables standard classifiers to use small datasets to achieve relatively high predictability by engineering new structured datasets encoded with temporal features, sampling strategies, and dimension reduction techniques, providing clinically useful prognostic models and setting an example for applying machine learning methods to small data problems in medicine.
Background and Objectives: We aim to establish deep learning models to optimize the individualized energy delivery for septic patients. Methods and Study Design: We conducted a study of adult septic patients in Intensive Care Unit (ICU), collecting 47 indicators for 14 days. After data cleaning and preprocessing, we used stats to explore energy delivery in deceased and surviving patients. We filtered out nutrition-related features and divided the data into three metabolic phases: acute early, acute late, and rehabilitation. Models were built using data before September 2020 and validated on the rest. We then established optimal energy target models for each phase using deep learning. Results: A total of 277 patients and 3115 data were included in this study. The models indicated that the optimal energy targets in the three phases were 900kcal/d, 2300kcal/d, and 2000kcal/d, respectively. Excessive energy intake increased mortality rapidly in the early period of the acute phase. Insufficient energy in the late period of the acute phase significantly raised the mortality of septic patients. For the rehabilitation phase, too much or too little energy delivery both associated with high mortality. Conclusion: Our study established time-series prediction models for septic patients to optimize energy delivery in the ICU. This approach indicated the feasibility of developing nutritional tools for critically ill patients. We recommended permissive underfeeding only in the early acute phase. Later, increased energy intake may improve survival and settle energy debts caused by underfeeding.
Extracting natural active ingredients from plants is an effective way to develop and screen modern drugs. Psoralea corylifolia is a leguminous plant whose seeds have long been used as a Traditional Chinese Medicine to treat psoriasis, rheumatism, dermatitis, and other diseases. To date, several main compounds, including coumarins, flavonoids, monoterpene phenols, and benzofurans, have been identified from the seeds of Psoralea corylifolia. Among them, bavachinin is a type of flavonoid with various biological activities. In this paper, the biological activities and mechanisms of action of bavachinin and its derivatives are reviewed. It includes the pharmacokinetic characteristics of bavachinin and its derivatives, as well as its prominent anti‐inflammatory, antitumor, antibacterial, and antiviral pharmacological activities and related metabolic studies. Bavachinin displayed these activities through different receptors, such as PPARs, as well as multiple signaling pathways and enzyme systems. In summary, bavachinin and its derivatives have potential drug development value in many fields, such as anti‐inflammatory, antitumor, nervous system disease, and diabetes. We believe that this review will lay a foundation for bavachinin‐based drug development throughout the world.
To investigate the predictive value of the arterial blood lactate to serum albumin ratio (LAR) on in-hospital mortality of patients with community-acquired pneumonia (CAP) admitted to the Intensive Care Unit (ICU).Clinical datasets of 1720 CAP patients admitted to ICU from MIMIC-IV database were retrospectively analyzed. Patients were randomly assigned to the training cohort (n=1204) and the validation cohort (n=516) in a ratio of 7:3. X-tile software was used to find the optimal cut-off value for LAR. The receiver operating curve (ROC) analysis was conducted to compare the performance between LAR and other indicators. Univariate and multivariate Cox regression analyses were applied to select prognostic factors associated with in-hospital mortality. Based on the observed prognostic factors, a nomogram model was created in training cohort, and the validation cohort was utilized to further validate the nomogram.The optimal cut-off value for LAR in CAP patients admitted to ICU was 1.6 (the units of lactate and albumin were, respectively, 'mmol/L' and 'g/dL'). The ROC analysis showed that the discrimination abilities of LAR were superior to other indicators except Sequential Organ Failure Assessment score and Simplified acute physiology score (SAPSII), which had the same abilities. Age, mean arterial pressure, SpO2, heart rate, SAPSII score, neutrophil-to-lymphocyte ratio, and LAR were found to be independent predictors of poor overall survival in the training cohort by multivariate Cox regression analysis and were incorporated into the nomogram for in-hospital mortality as independent factors. The nomogram model, exhibiting medium discrimination, had a C-index of 0.746 (95% CI = 0.715-0.777) in the training cohort and 0.716 (95% CI = 0.667-0.765) in the validation cohort.LAR could predict in-hospital mortality of patients with CAP admitted to ICU independently as a readily accessible biomarker. The nomogram that included LAR with other independent factors performed well in predicting in-hospital mortality.
Abstract Background Several studies have shown an association between plasma homocysteine levels and chronic obstructive pulmonary disease (COPD). It is not clear whether there is a causal association. A two-sample Mendelian randomization (MR) based method was used to further explore the causal association between plasma homocysteine and COPD. Methods Several studies have shown an association between plasma homocysteine levels and COPD. It is not clear whether there is a causal association.we performed a second data analysis using pooled data from published genome-wide association studies (GWASs) .we used genome-wide meta-analysis (n = 44147) to obtain genome-wide single nucleotide polypeptides (SNPs) associated with plasma homocysteine levels as instrumental variables. We used two-sample MR to study plasma homocysteine and COPD and COPD related diseases. MR analysis was performed by the random effects inverse variance weighting method and heterogeneity tests and pleiotropy tests were performed to evaluate the robustness of our findings. Results By two-sample MR analysis, We did not find causal associations between genetically predicted plasma homocysteine levels and COPD and COPD related diseases. In COPD hospital admissions,(OR = 1.06,95%CI 0.91–1.24, P = 0.42),asthma/COPD,(OR = 0.97,95%CI 0.89–1.06, P = 0.55),COPD related to chronic (opportunist) infection(OR = 1.50,95%CI 0.57–3.99, P = 0.41),COPD/asthma/ILD-related pneumonia or pneumonia-derived (OR = 0.93,95%CI 0.86–1.02, P = 0.13),COPD-related respiratory insufficiency(OR = 1.00,95%CI 0.7–1.44, P = 0.99), no heterogeneity and horizontal pleiotropy werefound. Conclusions Our study shows that genetically predicted plasma homocysteine levels are not causally associated with COPD, contrary to previous observational findings.As homocysteine is known to have deleterious effects on endothelial function and vascular homeostasis, further studies are needed to investigate whether additional factors mediate the association between homocysteine and COPD.
Abstract CA is a plant derivative with antibacterial and antiviral pharmacological effects, however, the therapeutic effect of CA on Klebsiella pneumonia and its mechanism study is still unclear. A rat KP model was established in vitro, a pneumonia cell model was established in vivo, the histopathological changes in the lungs were observed by HE staining after CA treatment, the expression of relevant inflammatory factors was detected by ELISA, the changes in the expression of proteins related to the AhR‐Src‐STAT3‐IL‐10 signaling pathway were detected by Western blot and immunofluorescence in the lungs, and the interactions between the proteins were verified by COIP relationship. The results showed that CA was able to attenuate the injury and inflammatory response of lung tissues, and molecular docking showed that there were binding sites between CA and AhR, and COIP demonstrated that AhR interacted with both STAT3 and Ser. In addition, CA was able to up‐regulate the expression levels of pathway‐related proteins of AhR, IL‐10, p‐Src, and p‐STAT3, and AhR knockdown was able to reduce LPS‐induced inflammatory responses and up‐regulate pathway‐related proteins, whereas CA treatment of AhR‐knockdown‐treated A549 cells did not show any statistically significant difference compared with the AhR knockdown group, demonstrating that CA exerts its pharmacological effects. These findings elucidated the mechanism of CA in the treatment of KP and demonstrated that CA is a potential therapeutic agent for KP.
Sivelestat sodium (SIV), a neutrophil elastase inhibitor, is mainly used for the clinical treatment of acute respiratory distress syndrome (ARDS) or acute lung injury (ALI). However, studies investigating the effects of SIV treatment of ALI are limited. Therefore, this study investigated the potential molecular mechanism of the protective effects of SIV against ALI. Human pulmonary microvascular endothelial cells (HPMECs) were stimulated with tumor necrosis factor α (TNF-α), and male Sprague-Dawley rats were intratracheally injected with Klebsiella pneumoniae (KP) and treated with SIV, ML385, and anisomycin (ANI) to mimic the pathogenetic process of ALI in vitro and in vivo, respectively. The levels of inflammatory cytokines and indicators of oxidative stress were assessed in vitro and in vivo. The wet/dry (W/D) ratio of lung tissues, histopathological changes, inflammatory cells levels in bronchoalveolar lavage fluid (BALF), and survival rates of rats were analyzed. The JNK/NF-κB (p65) and Nrf2/HO-1 levels in the HPMECs and lung tissues were analyzed by western blot and immunofluorescence analyses. Administration of SIV reduced the inflammatory factors levels, intracellular reactive oxygen species (ROS) production, and malondialdehyde (MDA) levels and increased the levels of superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px) in lung tissues. Meanwhile, SIV alleviated pathological injuries, decreased the W/D ratio, and inflammatory cell infiltration in lung tissue. In addition, SIV also inhibited the activation of JNK/NF-κB signaling pathway, promoted nuclear translocation of Nrf2, and upregulated the expression of heme oxygenase 1 (HO-1). However, ANI or ML385 significantly reversed these changes. SIV effectively attenuated the inflammatory response and oxidative stress. Its potential molecular mechanism was related to the JNK/NF-κB activation and Nrf2/HO-1 signaling pathway inhibition. This further deepened the understanding of the protective effects of SIV against ALI.
Sivelestat sodium (SIV), a neutrophil elastase inhibitor, is mainly used for the clinical treatment of acute respiratory distress syndrome (ARDS) or acute lung injury (ALI). However, studies investigating the effects of SIV treatment of ALI are limited. Therefore, this study investigated the potential molecular mechanism of the protective effects of SIV against ALI. Human pulmonary microvascular endothelial cells (HPMECs) were stimulated with tumor necrosis factor α (TNF-α), and male Sprague-Dawley rats were intratracheally injected with Klebsiella pneumoniae (KP) and treated with SIV, ML385, and anisomycin (ANI) to mimic the pathogenetic process of ALI in vitro and in vivo, respectively. The levels of inflammatory cytokines and indicators of oxidative stress were assessed in vitro and in vivo. The wet/dry (W/D) ratio of lung tissues, histopathological changes, inflammatory cells levels in bronchoalveolar lavage fluid (BALF), and survival rates of rats were analyzed. The JNK/NF-κB (p65) and Nrf2/HO-1 levels in the HPMECs and lung tissues were analyzed by western blot and immunofluorescence analyses. Administration of SIV reduced the inflammatory factors levels, intracellular reactive oxygen species (ROS) production, and malondialdehyde (MDA) levels and increased the levels of superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px) in lung tissues. Meanwhile, SIV alleviated pathological injuries, decreased the W/D ratio, and inflammatory cell infiltration in lung tissue. In addition, SIV also inhibited the activation of JNK/NF-κB signaling pathway, promoted nuclear translocation of Nrf2, and upregulated the expression of heme oxygenase 1 (HO-1). However, ANI or ML385 significantly reversed these changes. SIV effectively attenuated the inflammatory response and oxidative stress. Its potential molecular mechanism was related to the JNK/NF-κB activation and Nrf2/HO-1 signaling pathway inhibition. This further deepened the understanding of the protective effects of SIV against ALI.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)