Predicting delirium and the effects of medications in hospitalized COVID-19 patients using machine learning: A retrospective study within the Korean Multidisciplinary Cohort for Delirium Prevention (KoMCoDe).

Objective: Delirium is commonly reported from the inpatients with Coronavirus disease 2019 (COVID-19) infection. As delirium is closely associated with adverse clinical outcomes, prediction and prevention of delirium is critical. We developed a machine learning (ML) model to predict delirium in hospitalized patients with COVID-19 and to identify modifiable factors to prevent delirium. Methods: The data set (n = 878) from four medical centers was constructed. Total of 78 predictors were included such as demographic characteristics, vital signs, laboratory results and medication, and the primary outcome was delirium occurrence during hospitalization. For analysis, the extreme gradient boosting (XGBoost) algorithm was applied, and the most influential factors were selected by recursive feature elimination. Among the indicators of performance for ML model, the area under the curve of the receiver operating characteristic (AUROC) curve was selected as the evaluation metric. Results: Regarding the performance of developed delirium prediction model, the accuracy, precision, recall, F1 score, and the AUROC were calculated (0.944, 0.581, 0.421, 0.485, 0.873, respectively). The influential factors of delirium in this model included were mechanical ventilation, medication (antipsychotics, sedatives, ambroxol, piperacillin/tazobactam, acetaminophen, ceftriaxone, and propacetamol), and sodium ion concentration (all p < 0.05). Conclusions: We developed and internally validated an ML model to predict delirium in COVID-19 inpatients. The model identified modifiable factors associated with the development of delirium and could be clinically useful for the prediction and prevention of delirium in COVID-19 inpatients.

Proteome Analysis of the Hypothalamic Arcuate Nucleus in Chronic High-Fat Diet-Induced Obesity.

The hypothalamus plays a central role in the integrated regulation of feeding and energy homeostasis. The hypothalamic arcuate nucleus (ARC) contains a population of neurons that express orexigenic and anorexigenic factors and is thought to control feeding behavior via several neuronal circuits. In this study, a comparative proteomic analysis of low-fat control diet- (LFD-) and high-fat diet- (HFD-) induced hypothalamic ARC was performed to identify differentially expressed proteins (DEPs) related to changes in body weight. In the ARC in the hypothalamus, 6621 proteins (FDR < 0.01) were detected, and 178 proteins were categorized as DEPs (89 upregulated and 89 downregulated in the HFD group). Among the Gene Ontology molecular function terms associated with the DEPs, protein binding was the most significant. Fibroblast growth factor receptor substrate 2 (Frs2) and SHC adaptor protein 3 (Shc3) were related to protein binding and involved in the neurotrophin signaling pathway according to Kyoto Encyclopedia of Genes and Genomes analysis. Furthermore, high-precision quantitative proteomic analysis revealed that the protein profile of the ARC in mice with HFD-induced obesity differed from that in LFD mice, thereby offering insight into the molecular basis of feeding regulation and suggesting Frs2 and Shc3 as novel treatment targets for central anorexigenic signal induction.

Analysis of the determining factors of detectable P-wave and amplitude of QRS complex sensed by implantable loop recorder.

BACKGROUND: Determining factors for sufficient QRS amplitude and discernible P-wave sensing in implantable loop recorder (ILR) are unknown. We aimed to investigate determining factors and ILR implantation angle that may improve QRS complex and P-wave sensing in ILR. METHODS: We retrospectively reviewed 220 patients who underwent ILR implantation or follow-up analysis. Patient demographic, clinical, echocardiography, electrocardiography, heart angle, and ILR angle data were collected as predictor variables. Associations between ILR QRS amplitude/P-wave detectability and each predictor variable were investigated. RESULTS: Univariate linear regression showed that ILR QRS amplitude was significantly associated with age, height, ILR angle, and QRS amplitudes of 12-lead electrocardiogram (ECG) (lead I, II, aVR [inverted aVR], aVF, V1-V6) and Holter ECG (lead V3, V5). Among discrete variables, only left ventricular hypertrophy (LVH) affected ILR QRS amplitude (P = .016). A multivariate linear regression analysis revealed that ILR angle (ß = -0.008, P < .001), lead aVR amplitude (ß = 0.469, P = .003), Holter lead V5 amplitude (ß = 0.116, P = .049), Age (ß = -0.005, P = .014), and LVH (ß = 0.213, P = .031) were independent determinants of ILR QRS amplitude. Logistic regression revealed that heart angle significantly affected ILR P-wave detectability (ß = 0.12, P = .008). Multiple logistic regression revealed that heart angle (ß = 0.121, P = .013) and lead V1 amplitude (ß = 28.1, P = .034) were independent determinants of ILR P-wave detectability. CONCLUSION: ILR insertion angle, lead aVR QRS amplitude, Holter lead V5 QRS amplitude, age, and LVH are determinants of ILR QRS amplitude. Heart angle and lead V1 P-wave amplitude of 12-lead ECG are determinants of ILR P-wave detectability.

Optimizing tissue clearing and imaging methods for human brain tissue.

OBJECTIVES: To identify optimum sample conditions for human brains, we compared the clearing efficiency, antibody staining efficiency, and artifacts between fresh and cadaver samples. METHODS: Fresh and cadaver samples were cleared using X-CLARITY™. Clearing efficiency and artifact levels were calculated using ImageJ, and antibody staining efficiency was evaluated after confocal microscopy imaging. Three staining methods were compared: 4-day staining (4DS), 11-day staining (11DS), and 4-day staining with a commercial kit (4DS-C). The optimum staining method was then selected by evaluating staining time, depth, method complexity, contamination, and cost. RESULTS: Fresh samples outperformed cadaver samples in terms of the time and quality of clearing, artifacts, and 4',6-diamidino-2-phenylindole (DAPI) staining efficiency, but had a glial fibrillary acidic protein (GFAP) staining efficiency that was similar to that of cadaver samples. The penetration depth and DAPI staining improved in fresh samples as the incubation period lengthened. 4DS-C was the best method, with the deepest penetration. Human brain images containing blood vessels, cell nuclei, and astrocytes were visualized three-dimensionally. The chemical dye staining depth reached 800 µm and immunostaining depth exceeded 200 µm in 4 days. CONCLUSIONS: We present optimized sample preparation and staining protocols for the visualization of three-dimensional macrostructure in the human brain.

Radiologic Changes of Operated and Adjacent Segments after Anterior Cervical Microforaminotomy.

OBJECTIVE: Anterior cervical microforaminotomy (ACMF) is a motion-preserving surgical procedure. The purpose of this study is to assess radiologic changes of operated and adjacent segments after ACMF. METHODS: We retrospectively reviewed 52 patients who underwent ACMF between 1998 and 2008. From X-ray film-based changes, disc height and sagittal range of motion (ROM) of operated and adjacent segments were compared at preoperative and last follow-up periods. Radiological degeneration of both segments was analyzed as well. RESULTS: The mean follow-up period was 48.2 months. There were 78 operated, 52 upper adjacent, and 38 lower adjacent segments. There were statistically significant differences in the ROM and disc height of operated segment between preoperative and last follow-up periods. However, there were no statistically significant differences in the ROM and disc height of adjacent segment between both periods. Radiological degenerative changes of operated segments were observed in 30%. That of adjacent segments was observed in 11 and 11% at upper and lower segments, respectively. CONCLUSION: After mean 4-year follow-up periods, there were degenerative changes of operated segments. However, ACMF preserved motion and prevented degenerative changes of adjacent segments.