Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

ObjectiveTo investigate the contamination status, transmission risk and drug resistance of Klebsiella pneumoniae (KP) on the object surfaces in the surrounding environment of hospitalized patients infected with carbapenem-resistant Klebsiella pneumoniae (CRKP) , so as to provide a scientific guidance for the prevention and control of healthcare-associated infection. MethodsSamples from the surfaces of objects in the surrounding environment of CRKP infected patients living in the intensive care unit (ICU) and hand specimens from healthcare workers were collected for KP isolation and identification, as well as drug susceptible test in a medical institution located in Minhang District, Shanghai from 2021 to 2023. Additionally, both univariate and multivariate logistic regression analyses were used to identify the influencing factors associated with KP contamination in the hospital environment. ResultsA total of 546 surface samples were collected from the surrounding environment objects of 15 patients infected with CRKP, with a KP detection rate of 6.59% (36/546).The KP detection rate in the ICU of general ward (10.22%) was higher than that in the ICU of emergency department (2.94%) (χ²=12.142, P<0.001). Moreover, the KP detection rate on the surfaces of patient-contacted items (15.66%) was higher than that on shared-use items (6.25%), cleaning items (10.00%), and medical supplies (3.30%) (χ²=17.943, P<0.001). Besides, the detection rate of KP in items sent out of hospital for disinfection (15.38%) was higher than that in those self-disinfected (4.20%) (χ²=19.996, P<0.001).The highest detection rate of KP was observed in high-temperature washing (15.13%, 18/119) (χ²=21.219, P<0.001), while the lowest detection rate was observed in antibacterial hand sanitizer with trichlorohydroxydiphenyl ether sanitizing factor (0, 0/60) ( χ²=21.219, P<0.001).The detection rate of KP in samples taken more than 24 hours after the last disinfection (23.08%) was higher than that in those taken at 4 to24 hours (12.90%) and less than 4 hours (4.22%) (χ²=23.398,P<0.001).ICU of general ward (OR=4.045, 95%CI: 2.206‒7.416), patient-contacted items (OR=3.113, 95%CI: 1.191‒8.141), and self-disinfection ( OR=0.241, 95%CI:0.144‒0.402) were influencing factors for KP contamination in environmental surface. From 2021 to 2023, the drug resistance rates of hospital environmental KP isolates showed an upward trend (P<0.001) to antibiotics such as ceftazidime and gentamicin. Furthermore, high drug resistance rates of KP (>90%) were observed to ciprofloxacin, levofloxacin, cefotaxime, ceftriaxone, and cefepime. ConclusionCRKP can be transmitted outward through the surfaces of objects in the patients’ surroundings, and the drug resistance situation is severe. In clinical settings, it is necessary to implement isolation measures for CRKP infection patients, to increase the frequency of disinfection for objects in their surroundings, to strengthen hand hygiene practices, and to use antibiotics appropriately.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Peyronie's disease (PD) is a disorder characterized by fibrous plaque formation in the penile tissue that leads to curvature and complications in advanced stages. In this study, we aimed to compare four injectable induction agents for the establishment of a robust rat model of PD: transforming growth factor-β1 (TGF-β1), fibrin, sodium tetradecyl sulfate (STS) combined with TGF-β1, and polidocanol (POL) combined with TGF-β1. The results showed that injection of TGF-β1 or fibrin into the tunica albuginea induced pathological endpoints without causing penile curvature. The STS + TGF-β1 combination resulted in both histological and morphological alterations, but with a high incidence of localized necrosis that led to animal death. The POL + TGF-β1 combination produced pathological changes and curvature comparable to STS + TGF-β1 and led to fewer complications. In conclusion, fibrin, STS + TGF-β1, and POL + TGF-β1 all induced PD with a certain degree of penile curvature and histological fibrosis in rats. The POL + TGF-β1 combination offered comparatively greater safety and clinical relevance and may have the greatest potential for PD research using model rats.
Animals ; Male ; Penile Induration/drug therapy* ; Rats ; Transforming Growth Factor beta1/metabolism* ; Disease Models, Animal ; Fibrin ; Penis/drug effects* ; Polidocanol/administration & dosage* ; Rats, Sprague-Dawley ; Polyethylene Glycols/administration & dosage* ; Injections

The patient, an 11-year-old girl, was admitted with recurrent fever for 20 days, worsening with abdominal distension for 7 days. Upon admission, she presented with recurrent fever, lymphadenopathy, hepatosplenomegaly, polyserositis, and multiple organ dysfunction. Lymph node pathology and clinical manifestations confirmed the diagnosis of idiopathic multicentric Castleman disease-TAFRO syndrome. Treatment with siltuximab combined with glucocorticoids was initiated, followed by maintenance therapy with tocilizumab. The patient is currently in complete clinical remission. Therefore, once a child is diagnosed with idiopathic multicentric Castleman disease -TAFRO syndrome, early use of siltuximab should be considered for rapid disease control, followed by tocilizumab for maintenance therapy.
Humans ; Castleman Disease/drug therapy* ; Child ; Antibodies, Monoclonal, Humanized/administration & dosage* ; Female ; Antibodies, Monoclonal/administration & dosage*

OBJECTIVES: To investigate the role of active glucose monitoring in preventing hypoglycemia during the perioperative period of gastrointestinal endoscopy in children with glycogen storage disease type Ⅰb (GSD-Ⅰb). METHODS: A retrospective analysis was performed for the clinical data of children with GSD-Ⅰb who were diagnosed and treated in Guangdong Provincial People's Hospital from June 2021 to August 2024. The effect of active glucose monitoring on hypoglycemic episodes during the perioperative period of gastrointestinal endoscopy was analyzed. RESULTS: A total of 14 children with GSD-Ⅰb were included, among whom there were 7 boys and 7 girls, with a mean age of 10.0 years. Among 34 hospitalizations, there were 15 cases of hypoglycemic episodes (44%), among which 6 symptomatic cases (1 case with blood glucose level of 1.6 mmol/L and 5 cases with blood glucose level of <1.1 mmol/L) occurred without active monitoring, while 9 asymptomatic cases (with blood glucose level of 1.2-3.9 mmol/L) were detected by active monitoring. The predisposing factors for hypoglycemic episodes included preoperative fasting (5 cases, 33%), delayed feeding (7 cases, 47%), vomiting (2 cases, 13%), and parental omission (1 case, 7%). Two children experienced two hypoglycemic episodes during the same period of hospitalization, and no child experienced subjective symptoms prior to hypoglycemic episodes. Treatment methods included nasogastric glucose administration (1 case, 7%), intravenous injection of glucose (14 cases, 93%), and continuous glucose infusion (4 cases, 27%). Blood glucose returned to 3.5-6.9 mmol/L within 10 minutes after intervention and remained normal after dietary resumption. CONCLUSIONS Active glucose monitoring during the perioperative period of gastrointestinal endoscopy can help to achieve early detection of hypoglycemic states in children with GSD-Ⅰb, prevent hypoglycemic episodes, and enhance precise diagnosis and treatment.
Humans ; Female ; Male ; Child ; Retrospective Studies ; Blood Glucose/analysis* ; Hypoglycemia/etiology* ; Glycogen Storage Disease Type I/blood* ; Endoscopy, Gastrointestinal ; Perioperative Period ; Child, Preschool ; Adolescent

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

The neurophysiological mechanisms by which exercise improves cognitive function have not been fully elucidated. A comprehensive and systematic review of current domestic and international neurophysiological evidence on exercise improving cognitive function was conducted from multiple perspectives. At the molecular level, exercise promotes nerve cell regeneration and synaptogenesis and maintains cellular development and homeostasis through the modulation of a variety of neurotrophic factors, receptor activity, neuropeptides, and monoamine neurotransmitters, and by decreasing the levels of inflammatory factors and other modulators of neuroplasticity. At the cellular level, exercise enhances neural activation and control and improves brain structure through nerve regeneration, synaptogenesis, improved glial cell function and angiogenesis. At the structural level of the brain, exercise promotes cognitive function by affecting white and gray matter volumes, neural activation and brain region connectivity, as well as increasing cerebral blood flow. This review elucidates how exercise improves the internal environment at the molecular level, promotes cell regeneration and functional differentiation, and enhances the brain structure and neural efficiency. It provides a comprehensive, multi-dimensional explanation of the neurophysiological mechanisms through which exercise promotes cognitive function.
Animals ; Humans ; Brain/physiology* ; Cognition/physiology* ; Exercise/physiology* ; Nerve Regeneration/physiology* ; Neuronal Plasticity/physiology*