重症监护病房脓毒症患者血糖轨迹与预后的关联
基于MIMIC-IV数据库纳入8 923例成人脓毒症患者,采用基于组的轨迹模型分析入住ICU 7天内血糖轨迹,识别出持续高水平组(10.0-11.1 mmol/L,2 009例)、中等水平组(7.2-8.0 mmol/L,3 631例)和正常低限组(5.6-6.7 mmol/L,3 283例)。Kaplan-Meier曲线显示28天和1年累积生存率均以持续高水平组最低、正常低限组最高。多因素Logistic回归调整后,以持续高水平组为参照,中等水平组和正常低限组的院内死亡OR分别为0.677(0.574-0.799)和0.467(0.385-0.557),ICU死亡OR分别为0.627(0.524-0.751)和0.390(0.316-0.480)。整个队列仅4例发生低血糖,不同血糖轨迹组间7天内低血糖发生率无统计学显著差异。
要点
- 脓毒症患者入住ICU后7天内持续血糖10.0-11.1 mmol/L者院内死亡率20.01%、ICU死亡率16.38%,为三组中最高。
- 以持续高水平组为参照,正常低限组(5.6-6.7 mmol/L)院内死亡OR 0.467(0.385-0.557)、ICU死亡OR 0.390(0.316-0.480),均P<0.001。
- 中等水平组(7.2-8.0 mmol/L)较持续高水平组院内死亡OR 0.677(0.574-0.799)、ICU死亡OR 0.627(0.524-0.751),均P<0.001。
- 28天和1年累积生存率均以持续高水平组最低、正常低限组最高(28天Log-rank χ²=105.950,1年χ²=62.186,均P<0.001)。
- 整个队列仅4例低血糖(中等水平组3例、正常低限组1例),不同血糖轨迹组7天内低血糖发生率无统计学显著差异。
适用人群与场景
- 基于MIMIC-IV数据库的成人脓毒症患者,且入住ICU 7天内每天至少有一次血糖测量。
- 纳入分析8 923例,男性5 274例(59.11%)、女性3 649例(40.89%),年龄68(57,78)岁。
诊断标准与关键阈值
- 采用基于组的轨迹模型(GBTM)分析入住ICU 7天内血糖轨迹,结合临床相关性最终确定3个轨迹组。
- 持续高水平组2 009例:血糖持续10.0-11.1 mmol/L(180-200 mg/dL)。
- 中等水平组3 631例:血糖持续7.2-8.0 mmol/L(130-145 mg/dL)。
- 正常低限组3 283例:血糖持续5.6-6.7 mmol/L(100-120 mg/dL)。
研究设计与关键结果
- 研究类型为基于MIMIC-IV数据库的队列研究,最终纳入8 923例成人脓毒症患者。
- 主要终点28天累积生存率:Kaplan-Meier曲线显示持续高水平组最低、正常低限组最高(Log-rank χ²=105.950,P<0.001)。
- 1年累积生存率同样以持续高水平组最低、正常低限组最高(Log-rank χ²=62.186,P<0.001)。
- 院内死亡率:正常低限组10.66%(350/3 283)、中等水平组14.51%(527/3 631)、持续高水平组20.01%(402/2 009)。
- ICU死亡率:正常低限组7.46%(245/3 283)、中等水平组10.80%(392/3 631)、持续高水平组16.38%(329/2 009)。
- 以持续高水平组为参照,中等水平组院内死亡OR 0.677(0.574-0.799)、正常低限组OR 0.467(0.385-0.557);ICU死亡中等水平组OR 0.627(0.524-0.751)、正常低限组OR 0.390(0.316-0.480),均P<0.001。
- 随血糖水平升高,合并2型糖尿病比例、血乳酸水平、血尿素氮水平及使用血管升压药比例逐渐增加(均P<0.05)。
- 整个队列仅4例发生低血糖:中等水平组3例、正常低限组1例;Firth回归分析显示不同血糖轨迹组间入住ICU 7天内低血糖发生率无统计学显著差异。
证据类型、等级与局限
- 证据类型:基于MIMIC-IV数据库的队列研究。
- 局限:原文未报告局限性。
对床旁实践的改变
- 作者认为入住ICU后血糖轨迹与脓毒症患者预后密切相关,持续高血糖者死亡率最高。
- 作者认为随着血糖水平降低死亡率逐渐下降,维持血糖在较低水平5.6-6.7 mmol/L(100-120 mg/dL)与最低死亡率相关,且不增加低血糖风险。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开
OBJECTIVE: To investigate the relationship between dynamic blood glucose trajectories within 7 days of intensive care unit (ICU) admission and the risks of death and hypoglycemia in patients with sepsis.
METHODS: Based on data from the Medical Information Mart for Intensive Care-IV (MIMIC-IV), adult patients with sepsis who had at least one blood glucose measurement per day within 7 days of ICU admission were selected as study subjects. Group-based trajectory modeling (GBTM) was used to analyze blood glucose trajectories within 7 days of ICU admission and determine the optimal number of subgroups. General patient information (gender, age, race, ICU type), disease status (disease severity scores, comorbidities), and monitoring, laboratory parameters, and treatments during the first 24 hours of ICU admission were collected. The primary outcome was the 28-day cumulative survival rate. Secondary outcomes included the 1-year cumulative survival rate, in-hospital mortality, ICU mortality, and the incidence of hypoglycemia. The clinical characteristics of patients with different glucose trajectories were analyzed. Kaplan-Meier survival curve was used to analyze the relationship between glucose trajectories and 28-day and 1-year cumulative survival. Multivariate Logistic regression analysis or Firth regression analysis was employed to explore the relationship between glucose trajectories and in-hospital death, ICU death, and the occurrence of hypoglycemia within 7 days of ICU admission.
RESULTS: A total of 8 923 patients were finally enrolled, with 5 274 males (59.11%) and 3 649 females (40.89%), and the age was 68 (57, 78) years. Based on GBTM analysis and clinical relevance, three glucose trajectory groups were ultimately identified. According to the dynamic blood glucose changes within 7 days of ICU admission: the sustained high-level group included 2 009 patients, characterized by persistent blood glucose levels of 10.0-11.1 mmol/L (180-200 mg/dL); the moderate-level group included 3 631 patients, characterized by persistent blood glucose levels of 7.2-8.0 mmol/L (130-145 mg/dL); the normal low-limit group included 3 283 patients, characterized by persistent blood glucose levels of 5.6-6.7 mmol/L (100-120 mg/dL). Analysis of clinical characteristics showed that as blood glucose levels increased, the proportion of patients with comorbid type 2 diabetes mellitus, blood lactic acid (Lac) levels, blood urea nitrogen (BUN) levels, and the proportion of using vasopressors increased gradually (all P<0.05). Kaplan-Meier survival curve showed that for both 28-day and 1-year survival, the cumulative survival rate was lowest in the sustained high-level group and highest in the normal low-limit group (28-day Log-rank test: χ 2=105.950, P<0.001; 1-year Log-rank test: χ 2=62.186, P<0.001). Multivariate Logistic regression analysis, after adjusting for relevant variables, revealed that as blood glucose levels increased, in-hospital mortality and ICU mortality gradually rised [in-hospital mortality rates for the normal low-limit, moderate-level, and sustained high-level groups were 10.66% (350/3 283), 14.51% (527/3 631), and 20.01% (402/2 009), respectively; ICU mortality rates were 7.46% (245/3 283), 10.80% (392/3 631), and 16.38% (329/2 009), respectively]. Using the sustained high-level group as reference, the in-hospital mortality and ICU mortality in the moderate-level and normal low-limit groups were reduced [in-hospital mortality: odds ratio (OR) and 95% confidence intervals (95%CI) were 0.677 (0.574-0.799) and 0.467 (0.385-0.557), respectively; ICU mortality: the OR and 95%CI were 0.627 (0.524-0.751) and 0.390 (0.316-0.480), respectively; all P<0.001]. Only 4 cases of hypoglycemia occurred in the entire cohort: 3 in the moderate-level group and 1 in the normal low-limit group. Firth regression analysis showed that there was no statistically significant difference in the incidence of hypoglycemia within 7 days of ICU admission among the different glucose trajectory groups.
CONCLUSIONS: Post-ICU admission blood glucose trajectory is closely associated with prognosis in patients with sepsis, with the highest mortality observed in patients with sustained hyperglycemia. Mortality gradually decreased as blood glucose levels decreased; maintaining blood glucose at a lower level of 5.6-6.7 mmol/L (100-120 mg/dL) was associated with the lowest mortality without increasing the risk of hypoglycemia.
参考来源
- Wang H, Wu J, Liu J, et al. [Association between blood glucose trajectory and prognosis in septic patients in intensive care unit]. Zhonghua Wei Zhong Bing Ji Jiu Yi Xue. 2026. PMID: 41833353. doi:10.3760/cma.j.cn121430-20241030-00525
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。