U−shaped association between the glycemic variability and prognosis in hemorrhagic stroke patients: a retrospective cohort study from the MIMIC-IV database DOI Creative Commons
Yuchen Liu, Hung‐Chun Fu, Yue Wang

et al.

Frontiers in Endocrinology, Journal Year: 2025, Volume and Issue: 16

Published: April 3, 2025

Background Elevated glycemic variability (GV) is commonly observed in intensive care unit (ICU) patients and has been associated with clinical outcomes. However, the relationship between GV prognosis ICU hemorrhagic stroke (HS) remains unclear. This study aims to investigate association short- long-term all-cause mortality. Methods Clinical data for were obtained from MIMIC-IV 3.1 database. was quantified using coefficient of variation (CV), calculated as ratio standard deviation mean blood glucose level. The outcomes analyzed Cox proportional hazards regression models. Additionally, restricted cubic spline (RCS) curves employed examine nonlinear Results A total 2,240 HS included this study. In fully adjusted models, RCS analyses revealed a U-shaped CV both mortality (P nonlinearity < 0.001 all outcomes). Two-piecewise models subsequently applied identify thresholds. thresholds ICU, during hospitalization, at 30, 90, 180 days determined be 0.14, 0.16, 0.155, respectively. These findings consistent sensitivity subgroup analyses. Conclusions patients, higher an increased risk Our suggest that stabilizing may improve patients.

Language: Английский

Role of Oxidative Balance Score in Staging and Mortality Risk of Cardiovascular-Kidney-Metabolic Syndrome: Insights from Traditional and Machine Learning Approaches DOI Creative Commons
Yang Chen, Shuang Wu, Hongyu Liu

et al.

Redox Biology, Journal Year: 2025, Volume and Issue: 81, P. 103588 - 103588

Published: March 7, 2025

To evaluate the roles of oxidative balance score (OBS) in staging and mortality risk cardiovascular-kidney-metabolic syndrome (CKM). Data this study were from National Health Nutrition Examination Survey 1999-2018. We performed cross-sectional analyses using multinomial logistic regression to investigate relationship between OBS CKM staging. Cox proportional hazards models used assess impact on outcomes patients. Additionally, mediation explore whether mediated relationships specific predictors (Life's Simple 7 [LS7], systemic immune-inflammation index [SII], frailty score) outcomes. Then, machine learning developed classify stages 3/4 predict all-cause mortality, with SHapley Additive exPlanations values interpret contribution components. 21,609 participants included (20,319 CKM, median [IQR] age: 52.0 [38.0-65.0] years, 54.3% male, follow-up: 9.4 [5.3-14.1] years). Lower quartiles associated advanced Moreover, lower related increased risk, compared Q4 (all-cause mortality: Q1: HR 1.31, 95% CI 1.18-1.46, Q2: 1.27, 1.14-1.42, Q3: 1.18, 1.06-1.32; cardiovascular 1.44, 1.16-1.79, 1.39, 1.11-1.74, 1.26, 1.01-1.57; non-cardiovascular 1.12-1.44, 1.23, 1.08-1.40, 1.16, 1.02-1.31), optimal stratification threshold for was 22. (ranging 4.25%-32.85 %) effects SII, LS7, scores light gradient boosting achieved highest performance predicting (area under curve: 0.905) 0.875). Cotinine while magnesium, vitamin B6, physical activity protective. This highlights as a tool emphasizing stress's role management.

Language: Английский

Citations

2
Prognostic value of the Charlson Comorbidity Index for mortality in critically ill patients with paralytic ileus and mortality prediction model using machine learning (Preprint) DOI
Qingsheng Yu, Hui Feng, Fuhai Zhou

et al.

Published: April 15, 2025

BACKGROUND The burden of paralytic ileus (PI) in the intensive care unit (ICU) remains high, and Charlson Comorbidity Index (CCI) is strongly associated with prognosis several acute chronic diseases. However, there no literature on clinical value CCI as a prognostic assessment tool for critically ill patients PI ICU. OBJECTIVE aim this study was to investigate relationship between PI. METHODS In study, data from Critical Care Medical Information Marketplace IV 2.2 database were used determine optimal cutoff predicting mortality using receiver operating characteristic (ROC) curves, evaluated Cox regression restricted cubic spline analysis. A machine learning (ML) prediction model then constructed predict hospital by combining other characteristics. RESULTS included 863 (median age 65.4 years [interquartile range 54.6-75.5 years], 66.6% male). ROC curve identified an cut-off 4.5 CCI. Multivariate analysis showed that compared lowest quartile, elevated levels more likely have (Q4: HR 2.447, 95% CI 1.210-4.951), 28-day 3. 891, 1.956-7.740) 90-day 3.994, 2.224-7.173) all-cause significantly levels; however, association ICU 1.892, 0.653-5.480) weak. Among 11 ML models, LightGBM performed best, internal validation results showing area under 0.811, G-mean 0.670, F1 score 0.895. CONCLUSIONS important predictor hospital, 28-day, PI, threshold 4.5. models including show high accuracy mortality, occupies position model. This suggests helps identify high-risk patients, supports decision making, improves prognosis. CLINICALTRIAL NO

Language: Английский

Citations

0
Prognostic value of SAPS II score for 28-day mortality in ICU patients with acute pulmonary embolism DOI Creative Commons
Peng Liu,

Yongkui Ren

International Journal of Cardiology, Journal Year: 2025, Volume and Issue: unknown, P. 133201 - 133201

Published: March 1, 2025

Language: Английский

Citations

0
U−shaped association between the glycemic variability and prognosis in hemorrhagic stroke patients: a retrospective cohort study from the MIMIC-IV database DOI Creative Commons
Yuchen Liu, Hung‐Chun Fu, Yue Wang

et al.

Frontiers in Endocrinology, Journal Year: 2025, Volume and Issue: 16

Published: April 3, 2025

Background Elevated glycemic variability (GV) is commonly observed in intensive care unit (ICU) patients and has been associated with clinical outcomes. However, the relationship between GV prognosis ICU hemorrhagic stroke (HS) remains unclear. This study aims to investigate association short- long-term all-cause mortality. Methods Clinical data for were obtained from MIMIC-IV 3.1 database. was quantified using coefficient of variation (CV), calculated as ratio standard deviation mean blood glucose level. The outcomes analyzed Cox proportional hazards regression models. Additionally, restricted cubic spline (RCS) curves employed examine nonlinear Results A total 2,240 HS included this study. In fully adjusted models, RCS analyses revealed a U-shaped CV both mortality (P nonlinearity < 0.001 all outcomes). Two-piecewise models subsequently applied identify thresholds. thresholds ICU, during hospitalization, at 30, 90, 180 days determined be 0.14, 0.16, 0.155, respectively. These findings consistent sensitivity subgroup analyses. Conclusions patients, higher an increased risk Our suggest that stabilizing may improve patients.

Language: Английский

Citations

0