基础医学与临床 ›› 2026, Vol. 46 ›› Issue (9): 1256-1261.doi: 10.16352/j.issn.1001-6325.2026.09.1256

• 临床研究 • 上一篇    下一篇

胰腺术后患者扫描式葡萄糖监测准确性的影响因素

兰岭1, 张乐1, 张越伦2, 申乐1*, 黄宇光1   

  1. 中国医学科学院北京协和医学院 北京协和医院 1.麻醉科;2.医学研究中心, 北京 100730
  • 收稿日期:2026-05-20 修回日期:2026-06-22 出版日期:2026-09-05 发布日期:2026-08-18
  • 通讯作者: *pumchshenle@163.com
  • 基金资助:
    雄安新区科技创新专项(2023XAGG0070);协和人才培育计划A类项目(ULJ02859);协和人才培育计划B类项目(UGG10539);首都卫生发展科研专项(首发2024-2-4015)

Interference factors affecting the accuracy of flash glucose monitoring after pancreatic surgery

LAN Ling1, ZHANG Le1, ZHANG Yuelun2, SHEN Le1*, HUANG Yuguang1   

  1. 1. Department of Anesthesiology; 2. Medical Research Center, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing 100730, China
  • Received:2026-05-20 Revised:2026-06-22 Online:2026-09-05 Published:2026-08-18
  • Contact: *pumchshenle@163.com

摘要: 目的 探究扫描式葡萄糖监测(FGM)系统在胰腺手术患者术后阶段数值准确性的影响因素。方法 回顾性分析2024年 3 月至 2025 年 3 月间于北京协和医院择期行胰腺手术的279例,最终纳入255例成年患者。所有患者术前1 d于左上臂背侧佩戴FGM传感器且术后至少进行2次床旁即时(POC)毛细血管血糖分析。参考血糖值为与FGM值同一时期的POC血糖值,配对方法为以POC血糖值时间点为基准点,查询Gplus系统中最临近该基准点时间点的FGM值进行配对。计算每位患者的配对血糖绝对差值均值。利用LASSO回归法进行变量筛选,基于Bootstrap重抽样评估筛选变量的出现频次,采用多元线性回归法探究变量与配对血糖值差值均值的相关性。结果 FGM值与POC血糖值差值均值的中位数为1.2 mmol/L。多元线性回归分析结果显示术后血糖最大值(β=0.082,95% CI[0.055,0.108],P<0.001)、术中晶体液输注量(输注每1 L, β=0.124,95% CI[0.054,0.194],P<0.001)、黄疸(β=0.215,95% CI[0.019,0.412],P=0.031)与年龄(每增加10岁,β=0.091,95% CI[0.031,0.150],P=0.003)与配对血糖值差值均值的自然对数线性相关,该模型调整后的决定系数Radj2=0.245。结论 年龄、黄疸、术中晶体液输注量及术后血糖最大值均为影响FGM术后阶段数值准确性的独立影响因素,上述指标水平的升高均与FGM监测偏差增大线性相关。

关键词: 扫描式葡萄糖监测, 血糖, 黄疸, 手术

Abstract: Objective To explore the factors affecting the numerical accuracy of flash glucose monitoring (FGM) system in patients after pancreatic surgery. Methods A retrospective analysis was conducted on 279 adult patients who underwent elective pancreatic surgery at Peking Union Medical College Hospital from March 2024 to March 2025. Finally, 255 cases were included. All patients wore FGM sensors on the dorsum of the left upper arm one day before surgery, and received point-of-care (POC) capillary blood glucose measurements at least twice postoperatively. The reference blood glucose values were contemporaneous POC blood glucose readings. Paired matching was performed by taking the time points of POC blood glucose detection as benchmarks, and retrieving the closest FGM readings in the Gplus system for pairing. The mean absolute difference of paired blood glucose values was calculated for each subject. LASSO regression was used for variable selection, and Bootstrap resampling was applied to evaluate the occurrence frequency of screened variables. Multiple linear regression analysis was adopted to investigate the correlation between relevant variables and the mean difference of paired blood glucose values. Results The median mean difference between FGM readings and POC blood glucose values was 1.2 mmol/L. Multiple linear regression analysis revealed that maximum postoperative blood glucose (β=0.082, 95% CI: 0.055-0.108, P<0.001), intraoperative crystalloid infusion (per 1L increment, β=0.124, 95% CI: 0.054-0.194, P<0.001), jaundice (β=0.215, 95% CI: 0.019-0.412, P<0.05) and age (per decade increase, β=0.091, 95% CI: 0.031-0.150, P<0.05) were linearly correlated with the natural logarithm of the mean paired blood glucose difference(Radj2=0.245). Conclusions Age, jaundice, intraoperative crystalloid infusion volume and maximum postoperative blood glucose are independent influencing factors for postoperative FGM accuracy. Elevations of these indicators are linearly correlated with greater FGM deviation.

Key words: flash glucose monitoring, blood glucose, jaundice, surgery

中图分类号: