Table 1_Association between glycemic variability and acute kidney injury incidence in patients with cerebral infarction: an analysis of the MIMIC-IV database.xlsx

IntroductionGlycemic variability (GV) is an increasingly important predictive indicator of vascular occlusion-related complications. Studies have demonstrated that a higher GV is associated with poor outcomes in patients with cerebral infarction (CI). The prognostic utility of GV in CI patients for predicting acute kidney injury (AKI) remains inadequately characterized. This investigation systematically examines the pathophysiological relationship between acute glycemic fluctuations and AKI development in CI populations, with particular emphasis on temporal patterns of glucose dysregulation.
MethodsThis retrospective cohort analysis utilized data from the MIMIC-IV database, categorizing CI patients into quartiles based on GV metrics. Primary outcomes included AKI incidence and renal replacement therapy (RRT) initiation, with in-hospital mortality designated as the secondary endpoint. Analytical methodologies employed Kaplan-Meier survival curves with log-rank testing, multivariable-adjusted Cox proportional hazards regression, and logistic regression modeling to evaluate GV-AKI associations while controlling for critical confounders.
ResultsThe analytical cohort comprised 3,343 critically ill individuals extracted from the MIMIC-IV database. Kaplan-Meier curve analysis demonstrated progressively elevated cumulative risks of AKI development, RRT requirement, and in-hospital mortality among individuals with heightened GV. Following multivariable adjustment, logistic regression models and Cox proportional hazards analyses confirmed GV as an independent predictor of AKI progression, RRT dependency, and mortality risk in cerebral infarction patients.
ConclusionThis investigation identifies GV as an independent prognostic determinant for AKI development in cerebral infarction patients. GV demonstrates clinical utility as a biomarker for stratifying AKI risk in this population.

【引言】血糖变异性（Glycemic variability，GV）是日益受到重视的血管阻塞相关并发症预测指标。已有研究证实，较高的GV与脑梗死（cerebral infarction，CI）患者的不良预后相关。但目前对于GV在CI患者中预测急性肾损伤（acute kidney injury，AKI）的预后价值，尚未得到充分阐明。本研究系统探讨了CI人群中急性血糖波动与AKI发生之间的病理生理学关联，重点关注血糖调节异常的时间模式。
【方法】本项回顾性队列分析采用MIMIC-IV数据库的数据，基于GV指标将CI患者分为四个四分位数组。主要结局指标包括AKI发生率与肾脏替代治疗（renal replacement therapy，RRT）启动情况，院内死亡率设为次要终点。分析方法采用Kaplan-Meier生存曲线结合对数秩检验、多变量校正的Cox比例风险回归以及逻辑回归模型，在控制关键混杂因素的前提下评估GV与AKI的关联。
【结果】本分析队列共纳入从MIMIC-IV数据库提取的3343名重症患者。Kaplan-Meier曲线分析显示，随着GV水平升高，患者AKI发生风险、RRT需求率以及院内死亡率的累积风险均呈进行性升高。经多变量校正后，逻辑回归模型与Cox比例风险分析证实，GV是CI患者AKI进展、RRT依赖以及死亡风险的独立预测因子。
【结论】本研究证实GV可作为CI患者AKI发生的独立预后决定因素，其具备作为该人群AKI风险分层生物标志物的临床应用价值。