切换至 "中华医学电子期刊资源库"

中华临床实验室管理电子杂志 ›› 2023, Vol. 11 ›› Issue (03) : 151 -157. doi: 10.3877/cma.j.issn.2095-5820.2023.03.005

调查研究

基于MIMIC-Ⅲ数据库的急性胰腺炎并发脓毒症风险预测模型构建与评价
姜巧, 张溱乐, 张艳玲, 余展鹏()   
  1. 510510 广东广州,广东三九脑科医院重症监护病区
    510120 广东广州,广州医科大学第一临床学院
    511180 广东广州,广州医科大学金域检验学院;510700 广东广州,广州医科大学附属第五医院检验科
  • 收稿日期:2023-03-29 出版日期:2023-08-28
  • 通信作者: 余展鹏
  • 基金资助:
    广州市卫生健康科技项目(20221A010071)

Construction and evaluation of risk prediction model of acute pancreatitis complicated with sepsis based on MIMIC-Ⅲ database

Qiao Jiang, Zhenle Zhang, Yanling Zhang, Zhanpeng Yu()   

  1. Intensive Care Unit of Guangdong 999 Brain Hospital, Guangzhou Guangdong 510510, China
    The First Clinical College of Guangzhou Medical University, Guangzhou Guangdong 510120, China
    KingMed School of Laboratory Medicine, Guangzhou Medical University, Guangzhou Guangdong 511180, China; Department of Clinical Laboratory, The Fifth Affiliated of Guangzhou Medical University, Guangzhou Guangdong 510700, China
  • Received:2023-03-29 Published:2023-08-28
  • Corresponding author: Zhanpeng Yu
引用本文:

姜巧, 张溱乐, 张艳玲, 余展鹏. 基于MIMIC-Ⅲ数据库的急性胰腺炎并发脓毒症风险预测模型构建与评价[J]. 中华临床实验室管理电子杂志, 2023, 11(03): 151-157.

Qiao Jiang, Zhenle Zhang, Yanling Zhang, Zhanpeng Yu. Construction and evaluation of risk prediction model of acute pancreatitis complicated with sepsis based on MIMIC-Ⅲ database[J]. Chinese Journal of Clinical Laboratory Management(Electronic Edition), 2023, 11(03): 151-157.

目的

分析急性胰腺炎患者并发脓毒症的危险因素,并建立急性胰腺炎患者并发脓毒症的风险预测模型。

方法

从重症监护医疗信息集市(MIMIC-Ⅲ)数据库中提取急性胰腺炎患者临床数据,使用Lasso回归分析筛选急性胰腺炎并发脓毒症的潜在危险因素,在此基础上,使用Logistic回归分析构建急性胰腺炎并发脓毒症的风险预测模型并绘制列线图预测模型,通过计算一致性指数(C-index)、绘制该模型的校准曲线、临床决策曲线及临床影响曲线,评价模型的预测能力与临床适用性。

结果

建立的急性胰腺炎并发脓毒症风险预测模型共纳入了包括临床体征及实验室检验指标在内的10个潜在危险因素,模型的C-index为0.800,重抽样后验证C-index为0.774,列线图校准曲线显示校准曲线与理想曲线有较好的一致性,此外临床决策曲线与影响曲线分析结果显示本模型有较好的临床适用性,急性胰腺炎患者使用此列线图模型能获得临床净收益。

结论

基于MIMIC-Ⅲ构建的急性胰腺炎并发脓毒症风险预测模型具有良好的临床实用性。

Objective

Analyze the risk factors of sepsis in patients with acute pancreatitis, and establish a risk prediction model of sepsis in patients with acute pancreatitis.

Methods

Clinical data of patients with acute pancreatitis was extracted from the Medical Information Mart for Intensive Care-Ⅲ (MIMIC-Ⅲ) database. Lasso regression analysis was used to screen potential risk factors for concurrent sepsis in acute pancreatitis. Based on this, logistic regression analysis was used to construct a risk prediction model for concurrent sepsis in acute pancreatitis and a calibration curve was plotted. The predictive ability and clinical applicability of the model were evaluated by calculating the C-index, plotting the calibration curve of the model, and analyzing the clinical decision curve and impact curve.

Results

A risk prediction model for concurrent sepsis in acute pancreatitis was established, which included 10 potential risk factors including clinical signs and laboratory test results. The model had a C-index of 0.800, with a validation C-index of 0.774 after resampling. The calibration curve showed good consistency with the ideal curve. In addition, the clinical decision curve and impact curve analysis showed that this model has good clinical applicability. Acute pancreatitis patients can benefit from using this calibration curve model to obtain clinical net benefits.

Conclusions

The risk prediction model of acute pancreatitis complicated with sepsis based on MIMIC-Ⅲ has good clinical practicability.

表1 两组急性胰腺炎患者的人口学及临床特征
变量名称 非并发脓毒症组(n=203) 并发脓毒症组(n=147) P
年龄/(
x¯
±s
56.67±16.15 58.68±15.36 0.237
性别/例(%) 0.299
94(46) 59(40)
109(54) 88(60)
最高体温/例(%) 0.004
≤38 ℃ 129(64) 70(48)
>38 ℃ 74(36) 77(52)
是否出现心动过速/例(%) 0.023
60(30) 27(18)
143(70) 120(82)
是否出现气促/例(%) 0.002
142(70) 78(53)
61(30) 69(47)
SOFA评分 3(1,5) 5(3.5,9) <0.001
白细胞计数/(×109/L) 14.6(10.65,19.05) 15.3(11.2,21.75) 0.144
是否贫血/例(%) 0.054
91(45) 50(34)
112(55) 97(66)
红细胞压积/(
x¯
±s
32.74±5.98 32.26±6.46 0.477
平均红细胞体积 91(87,95) 90(87,96) 0.736
红细胞分布宽度/例(%) 0.164
正常 173(85) 116(79)
增高 30(15) 31(21)
凝血酶原时间/s 14.2(13.3,15.6) 14.9(13.5,16.25) 0.005
国际标准化比值 1.3(1.2,1.48) 1.3(1.2,1.55) 0.020
活化部分凝血活酶时间/s 29.1(26.15,34.37) 31.9(27.75,36.9) <0.001
钠离子/(mmol/L) 141(138,144) 141(138.5,144) 0.603
钾离子/(mmol/L) 4.3(3.9,4.8) 4.4(4,5) 0.209
氯离子/(mmol/L) 105(101,108) 104(100,109) 0.694
碳酸氢根离子/(mmol/L) 22(19,25) 20(16,23) <0.001
血清总钙/(mg/dl) 7.8(7.3,8.3) 7.3(6.4,7.8) <0.001
血清磷/(mg/dl) 2.5(1.7,3.35) 2.6(1.9,3.5) 0.332
血清镁/(mg/dl) 1.7(1.5,1.9) 1.7(1.5,1.9) 0.104
血清白蛋白/(g/L) 29.27(27,33) 28(24,30.5) <0.001
血糖/(mg/dl) 146(114,207.5) 165(127.5,243) 0.020
血尿素氮/(mg/dl) 16(10,27) 27(18,42) <0.001
血淀粉酶/(IU/L) 436(130,596.5) 560.1(126,842) 0.051
脂肪酶/(IU/L) 595(134,1328) 632(83.5,1567) 0.711
碱性磷酸酶/(IU/L) 105(71,167.5) 97(64,148) 0.137
AST/ALT/例(%) <0.001
≤2 178(88) 105(71)
>2 25(12) 42(29)
图1 Lasso回归模型筛选变量注:1A.交叉验证曲线;1B.系数路径图。
表2 急性胰腺炎并发脓毒症危险因素
图2 急性胰腺炎并发脓毒症列线图预测模型
图3 急性胰腺炎并发脓毒症预测模型的校准曲线
图4 急性胰腺炎并发脓毒症预测模型的临床决策曲线与临床影响曲线注:4A. 临床决策曲线;黑色实线表示所有急性胰腺炎患者均未并发脓毒症的净获益率,灰色实线表示所有患者均并发脓毒症的净获益率,蓝色曲线为本研究模型的决策曲线;4B. 临床影响曲线。
1
MOURAD M M, EVANS R, KALIDINDI V, et al. Prophylactic antibiotics in acute pancreatitis: Endless debate[J]. Annals of the royal college of surgeons of England, 2017, 99(2): 107-112.
2
郭喆, 关键. 重症急性胰腺炎预防与阻断急诊专家共识[J]. 中国急救医学, 2022, 42(5): 369-379.
3
BONGAERTS G P, SEVERIJNEN R S. A reassessment of the PROPATRIA study and its implications for probiotic therapy[J]. Nature biotechnology, 2016, 34(1): 55-63.
4
JOHNSON A E, POLLARD T J, SHEN L, et al. MIMIC-Ⅲ, a freely accessible critical care database[J]. Scientific data, 2016, 3: 160035.
5
ANGUS D C, LINDE-ZWIRBLE W T, LIDICKER J,et al. Epidemiology of severe sepsis in the United States: Analysis of incidence, outcome, and associated costs of care[J]. Critical care medicine, 2001, 29(7): 1303-1310.
6
吴月, 孔婧, 刘宇洋, 等. 重症急性胰腺炎并发胰性脑病的诊疗进展[J]. 中国急救医学, 2023, 43(2): 156-160.
7
王天刚, 王星月, 喻玉, 等. 重症急性胰腺炎并发小肠扭转一例[J]. 中华消化病与影像杂志(电子版), 2022, 12(6): 377-378.
8
张春, 林婷, 张靖垚, 等. 急性胰腺炎合并区域性门静脉高压症诊治分析[J]. 中华肝脏外科手术学电子杂志, 2022, 11(1): 48-53.
9
PENG T, PENG X, HUANG M, et al. Serum calcium as an indicator of persistent organ failure in acute pancreatitis[J]. American journal of emergency medicine, 2017, 35(7): 978-982.
10
KEMPARAJ T, NARASIMHAMURTHY N, MURALIDHAR A. Total serum calcium and corrected calcium as a predictor of severity in acute pancreatitis[J]. International surgery journal, 2018, 5: 3558.
11
徐超, 刘君, 卢洪军, 等. 红细胞分布宽度、D-二聚体和血清钙离子水平对妊娠期急性胰腺炎病情严重程度的预测价值[J]. 热带医学杂志, 2021, 21(9): 1192-1196, 1217.
12
WALLER A, LONG B, KOYFMAN A, et al. Acute pancreatitis: Updates for emergency clinicians[J]. Journal of emergency medicine, 2018, 55(6): 769-779.
13
李娟, 朱向荣, 章霓, 等. 肠内营养对晚期胃癌化疗患者营养指标的影响[J]. 中国临床保健杂志, 2023, 26(2): 230-232.
14
KEMPKER J A, WANG H E, MARTIN G S. Sepsis is a preventable public health problem[J]. Critical care, 2018, 22: 116.
15
BEZAN A, MRSIC E, KRIEGER D, et al. The Preoperative AST/ALT (De Ritis) ratio represents a poor prognostic factor in a cohort of patients with nonmetastatic renal cell carcinoma[J]. The journal of urology, 2015, 194(1): 30-35.
16
MULLINS M E, MCGOVERN A J, et al. Can AST/ALT ratio indicate recovery after acute paracetamol poisoning? The authors reply. [J]. Clinical toxicology (Philadelphia, Pa.), 2015, 53(6): 581-582.
17
王庆春. 血清ALT、AST联合血脂检测对诊断脂肪肝的临床价值分析[J]. 现代诊断与治疗, 2022, 33(24): 3727-3729.
18
KOUTROUMPAKIS E, WU B U, BAKKER O J, et al. Admission hematocrit and rise in blood urea nitrogen at 24 h outperform other laboratory markers in predicting persistent organ failure and pancreatic necrosis in acute pancreatitis: A post hoc analysis of three large prospective databases[J]. American journal of gastroenterology, 2015, 110(12): 1707-1716.
19
张鸿, 邓川, 余佳珍, 等. 成年肾病综合征患者血清肌酐, 尿素氮与尿酸水平关联分析[J]. 公共卫生与预防医学, 2023, 34(1): 109-112.
20
ZHANG H, MENG F, LU S. Nomograms predicting the occurrence of sepsis in patients following major hepatobiliary and pancreatic surgery[J]. Gastroenterology research and practice, 2020:9761878.
21
GARCÍA-VILLALBA E, CANO-SÁNCHEZ A, ALCARAZ-GARCÍA A, et al. Nomogram to predict a poor outcome in emergency patients with sepsis and at low risk of organ damage according to sepsis-related organ failure assessment (SOFA)[J]. Emergencias: Revista de la sociedad española de medicina de emergencias, 2017, 29(2): 81-86.
22
舒欣, 李昊洋, 李雨捷, 等. 基于机器学习建立腹部手术术后脓毒症患者死亡风险预测模型[J]. 陆军军医大学学报, 2023, 45(8): 732-738.
23
卢苏, 孙健, 童朝阳. 乳酸与48 h乳酸清除率对脓毒症患者预后的预测价值[J]. 中国临床医学, 2023, 30(1): 91-96.
24
赵雪, 周坤鹏, 李鹏飞, 等. 早期平均血小板体积变化对脓毒症休克患者预后的预测价值[J]. 中华老年多器官疾病杂志, 2021, 20(5): 326-331.
25
袁波, 代华, 伍佳, 等. 人工智能在全科医学领域的应用[J].中华全科医学, 2021, 19(9): 1433-1436, 1572.
[1] 张秋彬, 张楠, 林清婷, 徐军, 朱华栋, 姜辉. 急性胰腺炎合并急性肾损伤患者的预后评估[J]. 中华危重症医学杂志(电子版), 2023, 16(05): 382-389.
[2] 陈永庄, 莫小乔, 谢天. 心血管事件患者术后30 d死亡风险决策树模型的构建与评估——基于少数类样本合成过采样技术算法[J]. 中华危重症医学杂志(电子版), 2023, 16(05): 390-398.
[3] 郑鹏, 吴赛萍, 谢秀璋, 史庆丰. 术前预测感染性肾结石列线图模型的构建及验证[J]. 中华实验和临床感染病杂志(电子版), 2023, 17(05): 299-306.
[4] 张俊, 罗再, 段茗玉, 裘正军, 黄陈. 胃癌预后预测模型的研究进展[J]. 中华普通外科学文献(电子版), 2023, 17(06): 456-461.
[5] 唐旭, 韩冰, 刘威, 陈茹星. 结直肠癌根治术后隐匿性肝转移危险因素分析及预测模型构建[J]. 中华普外科手术学杂志(电子版), 2024, 18(01): 16-20.
[6] 甄子铂, 刘金虎. 基于列线图模型探究静脉全身麻醉腹腔镜胆囊切除术患者术后肠道功能紊乱的影响因素[J]. 中华普外科手术学杂志(电子版), 2024, 18(01): 61-65.
[7] 王万川, 麦广智, 张晓槟, 冯家豪. 腹腔镜下LAR术治疗低位直肠癌保肛术中钉仓数量与术后LARS的关系[J]. 中华普外科手术学杂志(电子版), 2023, 17(05): 491-496.
[8] 贾成朋, 王代宏, 陈华, 孙备. 可切除性胰腺癌预后术前预测模型的建立及应用[J]. 中华普外科手术学杂志(电子版), 2023, 17(05): 566-570.
[9] 王晓丹, 王媛, 崔向宇, 任晓磊. 上尿路结石内镜手术后尿源性脓毒血症病原菌耐药及死亡高危因素分析[J]. 中华腔镜泌尿外科杂志(电子版), 2023, 17(06): 611-615.
[10] 李飞翔, 段虎斌, 李晋虎, 吴昊, 王永红, 范益民. 急性颅脑损伤继发下肢静脉血栓的相关危险因素分析及预测模型构建[J]. 中华神经创伤外科电子杂志, 2023, 09(05): 277-282.
[11] 王小红, 钱晶, 翁文俊, 周国雄, 朱顺星, 祁小鸣, 刘春, 王萍, 沈伟, 程睿智, 秦璟灏. 巯基丙酮酸硫基转移酶调控核因子κB信号介导自噬对重症急性胰腺炎大鼠的影响及机制[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 422-426.
[12] 陆萍, 邹健. 凝血和纤维蛋白溶解标志物的动态变化对急性胰腺炎患者预后的评估价值[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 427-432.
[13] 吉茜茜, 田尧, 马林, 钱进. 红细胞分布宽度-白蛋白比值联合BISAP评分对急性胰腺炎严重程度及死亡率的预测价值[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 433-438.
[14] 黄岩, 刘晓巍, 杨春玲, 兰烨. 急性胰腺炎合并糖尿病患者的临床特征及血糖代谢与病情严重度的相关性[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 439-442.
[15] 王亚丹, 吴静, 黄博洋, 王苗苗, 郭春梅, 宿慧, 王沧海, 王静, 丁鹏鹏, 刘红. 白光内镜下结直肠肿瘤性质预测模型的构建与验证[J]. 中华临床医师杂志(电子版), 2023, 17(06): 655-661.
阅读次数
全文


摘要