医学
内科学
纤维蛋白原
逻辑回归
败血症
队列研究
回顾性队列研究
队列
纵向研究
多元分析
生存分析
前瞻性队列研究
比例危险模型
心脏病学
试验预测值
流行病学
危险分层
观察研究
多元统计
子群分析
外科
疾病严重程度
联想(心理学)
作者
Jiamei Li,Ruohan Li,Yanli Hou,Yawen Yang,Qingmiao Ma,Xuting Jin,Jiajia Ren,J B Zhang,Ya Gao,Xi Zhang,Gang Wang
摘要
Longitudinal fibrinogen (FIB) trajectories in sepsis and their relationship with outcomes remain poorly understood. This study aimed to identify distinct FIB trajectories and their association with outcomes. This retrospective cohort study included 870 sepsis patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Generalized additive mixed model (GAMM) compared FIB trajectories between survivors and non-survivors. Group-based trajectory model (GBTM) was used to identify distinct trajectories. Logistic regression was performed to assess the association between trajectories and mortality. Sensitivity analysis was conducted using inverse probability of treatment weighting. Subgroup analysis was stratified by the sequential organ failure assessment score. The GAMM revealed consistently lower FIB level over time in non-survivors compared with survivors (P < 0.001), who exhibited a marked and sustained increase. Furthermore, the GBTM identified four distinct trajectories. Compared with Group 1 (persistent low; 37.5% mortality), Group 2 (moderate-rising to plateau; 20.55%; OR: 0.409, 95% CI: 0.268–0.624) and Group 3 (high-steady increase; 22.76%; 0.529, 0.326–0.859) had significantly lower ICU mortality, while Group 4 (very high-increasing; 27.78%; 0.571, 0.319–1.025) showed no significant survival benefit. These findings remained robust in sensitivity analysis. Longitudinal FIB trajectories were associated with outcomes. Persistently low FIB (<2.0 g/L) identified patients at the highest risk of mortality, whereas moderate elevations were associated with a survival benefit. Extreme hyperfibrinogenemia (>7.0 g/L) did not demonstrate a significant survival benefit. Monitoring FIB dynamics may aid in risk stratification and clinical decision-making in sepsis.
科研通智能强力驱动
Strongly Powered by AbleSci AI