Risk analysis of ventilator-associated pneumonia in patients with mechanical ventilation

doi:10.3877/cma.j.issn.1674-6902.2025.03.019

Abstract

Abstract:

Objective

To analyze the risk of ventilator-associated pneumonia （VAP） in patients with mechanical ventilation based on an interpretable machine learning model.

Methods

All of 77 patients with mechanical ventilation in our hospital from January 2022 to December 2024 were selected. 19 cases with VAP were divided into the observation group and 58 cases without VAP were divided into the control group. Clinical data were collected，and the features related to VAP were screened by LASSO regression analysis. Logistic regression （LR） and support vector machine were used. SVM，multi-layer perceptron （MLP） and eXtreme Gradient Boosting （xGBoost） were four machine learning（ML） algorithms to construct the prediction model，and SHAP was used to analyze the interpretability of the model.

Results

Univariate analysis showed that acute physiology and chronic health evaluation Ⅱ，APACHEⅡ）（21.53±2.50） minutes and mechanical ventilation time （10.37±3.30） d were higher than those of control group （18.59±2.80） minutes and mechanical ventilation time （8.38±2.55） d. Glasgow score （10.79±2.07） and albumin level （30.08±3.79） g/L in the observation group were lower than those in the control group （12.14±1.94） and albumin level （34.22±5.20） g/L. There were significant differences in age，chronic obstructive pulmonary disease，diabetes，sedative use and antibiotic type between observation group and control group （P ＜0.05）. LASSO regression results showed that age，APACHE Ⅱscore，mechanical ventilation time and albumin were risk factors for VAP in patients with mechanical ventilation. Four ML models were constructed based on the above four risk factors，and the area under the curve（AUC） of xGBoost model was 0.882，the accuracy was 81.30%，the sensitivity was 84.00%，and the specificity was 82.00%. SHAP results showed that the top 4 characteristics of VAP risk contribution were APACHE Ⅱscore，age，mechanical ventilation time and albumin.

Conclusion

xGBoost model has high sensitivity and specificity，which can be used as an auxiliary diagnostic tool to quickly and accurately determine the risk of VAP in patients with mechanical ventilation，and provide reference for clinical decision-making.

Key words: Ventilator-associated pneumonia, Mechanical ventilation, Machine learning, Risk analysis

Limei Xu, Haiyan Wu, Haili Lin, Shunni Ji, Chunru Chen, Jiaojiao Fu, Zhongren Chen, Xiaomei Wu. Risk analysis of ventilator-associated pneumonia in patients with mechanical ventilation[J]. Chinese Journal of Lung Diseases(Electronic Edition), 2025, 18(03): 447-451.

/ / Recommend

Add to citation manager EndNote|Ris|BibTeX

URL: https://zhfbjbzz.cma-cmc.com.cn/EN/10.3877/cma.j.issn.1674-6902.2025.03.019

https://zhfbjbzz.cma-cmc.com.cn/EN/Y2025/V18/I03/447

Figures/Tables 5

References 23

1	Trudzinski FC，Neetz B，Bornitz F，et al. Risk factors for prolonged mechanical ventilation and weaning failure：A systematic review［J］.Respiration，2022，101（10）：959-969.
2	Zorrilla-Vaca A，Arevalo JJ，Grant MC. Protective mechanical ventilation in critically ill patients after surgery［J］. Curr Opin Crit Care，2024，30（6）：679-683.
3	Tanios MA.Reconnecting to improve liberation from mechanical ventilation：A new perspective［J］. Chest，2024，165（6）：1282-1283.
4	金静怡，王丽华，季建红. ICU 呼吸机相关性肺炎风险预测模型的研究进展［J］. 当代护士，2024，31（22）：13-17.
5	张雪梅，董雪，郭燕杰. 呼吸机相关性肺炎风险预测模型的研究进展［J］. 全科护理，2022，20（32）：4497-4499.
6	李鹏，刘业，端另波，等. 俯卧位机械通气在重症肺炎患者中的疗效及并发呼吸机相关性肺炎风险因素的研究［J］. 大医生，2024，9（4）：109-112.
7	温慧，聂清梅，孙莉莉，等. ICU 机械通气患者呼吸机相关性肺炎风险预测模型的系统评价［J］. 中华现代护理杂志，2024，30（24）：3280-3286.
8	Martin-Loeches I，Torres A. Hospital-acquired pneumonia/ventilatorassociated pneumonia after guidelines［J］. Semin Respir Crit Care Med，2022，43（2）：173-174.
9	时铭蔚，李君，孙春萍，等. 基于机器学习算法构建老年人呼吸机相关肺炎早期预警模型［J］. 中华老年医学杂志，2023，42（6）：670-675.
10	燕红玲，王岩岩，陈树斌. PCT、NLR 联合LUBS 预测ICU CRKP致呼吸机相关肺炎撤机及预后分析［J/CD］. 中华肺部疾病杂志（电子版），2024，17（4）：617-620.
11	中华医学会重症医学分会. 机械通气临床应用指南（2006）［J］.中国危重病急救医学，2007，19（2）：65-72.
12	廖理粤，苏越明，邓方阁，等. 呼吸机相关性肺炎诊断、预防和治疗指南的质量评价［J］. 国际呼吸杂志，2016，36（14）：1041-1044.
13	Metersky ML，Kalil AC. Management of ventilator-associated pneumonia：Guidelines［J］. Infect Dis Clin North Am，2024，38（1）：87-101.
14	Howroyd F，Chacko C，MacDuff A，et al. Ventilator-associated pneumonia：pathobiological heterogeneity and diagnostic challenges［J］. Nat Commun，2024，15（1）：6447.
15	Safavi A，Molavynejad S，Rashidi M，et al. The effect of an infection control guideline on the incidence of ventilator-associated pneumonia in patients admitted to the intensive care units［J］. BMC Infect Dis，2023，23（1）：198.
16	王成立，周林裕，梁哲勇. 基于TSP-1 和APACHEⅡ评分构建老年重型颅脑损伤患者院内死亡的预测模型［J］. 中国实用神经疾病杂志，2024，27（7）：816-820.
17	林立伟，陈湘平，张金炎，等. APACHE Ⅱ评分与氧合指数对老年重症肺炎临床疗效及预后的评估价值［J］. 西藏医药，2024，45（1）：82-84.
18	Hassan ME，Al-Khawaja SA，Saeed NK，et al. Causative bacteria of ventilator-associated pneumonia in intensive care unit in Bahrain：Prevalence and antibiotics susceptibility pattern［J］. World J Crit Care Med，2023，12（3）：165-175.
19	Trudzinski FC，Neetz B，Bornitz F，et al. Risk factors for prolonged mechanical ventilation and weaning failure：A systematic review［J］.Respiration，2022，101（10）：959-969.
20	Foucrier A，Dessalle T，Tuffet S，et al. Association between combination antibiotic therapy as opposed as monotherapy and outcomes of ICU patients with Pseudomonas aeruginosa ventilator-associated pneumonia：an ancillary study of the iDIAPASON trial［J］. Crit Care，2023，27（1）：211.
21	Tang H，Yuan Z，Li J，et al. Fiberoptic bronchoscopy for the prevention of ventilator-associated pneumonia：a meta-analysis of randomized controlled trials［J］. J Infect Dev Ctries，2024，18（9）：1413-1420.
22	Xue-Meng C，Gao-Wang L，Xiao-Mei L，et al. Effect of mechanical ventilation under intubation on respiratory tract change of bacterial count and alteration of bacterial flora［J］. Exp Lung Res，2023，49（1）：165-177.
23	Lu P，Li D，Tian Q，et al. Effect of mixed probiotics on pulmonary flora in patients with mechanical ventilation：an exploratory randomized intervention study［J］. Eur J Med Res，2024，29（1）：473.

临床资料	观察组（n=19）	对照组（n=58）	t/χ²	P 值
临床资料	观察组（n=19）	对照组（n=58）	t/χ²	P 值	APACHE Ⅱ评分 [分,（x̄ ± s）]	21.53±2.50	18.59±2.80	4.316	0.000
格拉斯哥评分 [分,（x̄ ± s）]	10.79±2.07	12.14±1.94	2.505	0.018
吸烟[n(%)]	8(42.11)	22(37.93)	0.105	0.746
合并疾病[n(%)]
慢性阻塞性肺疾病	9(47.38)	10(17.24)	6.989	0.008
冠心病	5(26.32)	16(27.59)	0.012	0.914
高血压	6(31.58)	19(32.76)	0.009	0.924
糖尿病	9(47.38)	10(17.24)	6.989	0.008
白蛋白 [g/L,（x̄ ± s）]	30.08±3.79	34.22±5.20	3.745	0.001
C 反应蛋白 [mg/L,（x̄ ± s）]	7.89±2.12	7.64±2.14	0.445	0.659
胸部 CT 肺部感染征象[n(%)]	4(21.05)	10(17.24)	0.140	0.709
机械通气时间 [d,（x̄ ± s）]	10.37±3.30	8.38±2.55	2.404	0.024
使用镇静剂 [n(%)]	8(42.11)	9(15.52)	5.881	0.015
气管插管途径[n(%)]			0.037	0.847
气管切开	9(47.37)	26(44.83)
经口插管	10(52.63)	32(55.17)
应用抗生素种类 ≥3 种	8(42.11)	11(18.97)	4.123	0.042

临床资料	观察组（n=19）	对照组（n=58）	t/χ²	P 值
临床资料	观察组（n=19）	对照组（n=58）	t/χ²	P 值	APACHE Ⅱ评分 [分,（x̄ ± s）]	21.53±2.50	18.59±2.80	4.316	0.000
格拉斯哥评分 [分,（x̄ ± s）]	10.79±2.07	12.14±1.94	2.505	0.018
吸烟[n(%)]	8(42.11)	22(37.93)	0.105	0.746
合并疾病[n(%)]
慢性阻塞性肺疾病	9(47.38)	10(17.24)	6.989	0.008
冠心病	5(26.32)	16(27.59)	0.012	0.914
高血压	6(31.58)	19(32.76)	0.009	0.924
糖尿病	9(47.38)	10(17.24)	6.989	0.008
白蛋白 [g/L,（x̄ ± s）]	30.08±3.79	34.22±5.20	3.745	0.001
C 反应蛋白 [mg/L,（x̄ ± s）]	7.89±2.12	7.64±2.14	0.445	0.659
胸部 CT 肺部感染征象[n(%)]	4(21.05)	10(17.24)	0.140	0.709
机械通气时间 [d,（x̄ ± s）]	10.37±3.30	8.38±2.55	2.404	0.024
使用镇静剂 [n(%)]	8(42.11)	9(15.52)	5.881	0.015
气管插管途径[n(%)]			0.037	0.847
气管切开	9(47.37)	26(44.83)
经口插管	10(52.63)	32(55.17)
应用抗生素种类 ≥3 种	8(42.11)	11(18.97)	4.123	0.042

学习预测	AUC	敏感度	特异度	准确度	F1 分数	Brier 评分
LR	0.712	64.00%	71.00%	72.65%	0.671	0.320
SVM	0.725	70.00%	73.50%	75.00%	0.680	0.275
MLP	0.789	76.25%	78.50%	78.50%	0.768	0.251
XGboost	0.882	84.00%	82.00%	81.30%	0.833	0.196

学习预测	AUC	敏感度	特异度	准确度	F1 分数	Brier 评分
LR	0.712	64.00%	71.00%	72.65%	0.671	0.320
SVM	0.725	70.00%	73.50%	75.00%	0.680	0.275
MLP	0.789	76.25%	78.50%	78.50%	0.768	0.251
XGboost	0.882	84.00%	82.00%	81.30%	0.833	0.196

[1]	Xinyue Wang, Ya Yuan, Hua Shu, Kunpeng Cao, Xinhua Ye, Lu Li. Development of a dose prediction model for ultrasound-guided thrombin treatment of pseudoaneurysms using the LightGBM algorithm [J]. Chinese Journal of Medical Ultrasound (Electronic Edition), 2025, 22(02): 153-161.
[2]	Haoyu Fang, Xiao Wang, Anwei Zhang, Dandan Shang, Jiong Yu, Hongcui Cao. Biomarkers of mesenchymal stem cell therapy in Crohn's disease model mice based on fecal metabolomics [J]. Chinese Journal of Critical Care Medicine(Electronic Edition), 2025, 18(02): 98-104.
[3]	Yueqiu Shen, Menglin Cao, Meijie Xu, Mindan Wu, Kaiyi Wu, Linjuan Lu. Risk research of acute exacerbation in patients with chronic obstructive pulmonary disease based on interpretable machine learning model [J]. Chinese Journal of Lung Diseases(Electronic Edition), 2025, 18(03): 380-384.
[4]	Heng Zhang, Ge Shi, Ze Zhang, Zhenxuan Gao, Wenqiang Yang, Qi Wang, Li Zhang. Establishment of biomarkers for diabetic peripheral neuropathy based on plasma transcriptomics [J]. Chinese Journal of Neurotraumatic Surgery(Electronic Edition), 2025, 11(02): 127-134.
[5]	Guyi Wang, Jiaqi Li, Yanjun Zhong, Bo Yu, Chenfang Wu, Haiyun Dong, Min Xu, Huaqin Wang, Li Tang, Yanyan Zhu, Jinxiu Li, Ben Lyu. Impact of ARDS management strategy based on smart ICU platform on clinical outcome of ARDS patients [J]. Chinese Journal of Critical Care & Intensive Care Medicine(Electronic Edition), 2025, 11(01): 72-77.
[6]	Yunhui Ni, Xueyan Yuan, Lili Huang, Yi Yang, Haibo Qiu. Research progress in respiratory drive in protective ventilation strategy [J]. Chinese Journal of Critical Care & Intensive Care Medicine(Electronic Edition), 2025, 11(01): 91-94.
[7]	Qing Pan, Huiqing Ge. Review on automatic detection methods of patient-ventilatory asynchrony based on big data of mechanical ventilation waveform [J]. Chinese Journal of Critical Care & Intensive Care Medicine(Electronic Edition), 2024, 10(04): 399-403.
[8]	Mingyue Miao, Jianxin Zhou. Lung protective sedation: emphasize bedside assessment of respiratory drive and inspiratory effort [J]. Chinese Journal of Critical Care & Intensive Care Medicine(Electronic Edition), 2024, 10(04): 325-328.
[9]	Yahui Li, Lin Luan, Huiyun Huang. Construction and validation of risk model for young and middle-aged patients with liver cancer at different recurrence stages after radical operation [J]. Chinese Journal of Digestion and Medical Imageology(Electronic Edition), 2025, 15(02): 112-119.
[10]	Jian Chen, Kaijian Xia, Lu Xu, Fuli Gao, Xiaodan Xu, Ganhong Wang. Development of a risk assessment model for acute kidney injury in cirrhotic patients based on automated machine learning [J]. Chinese Journal of Clinicians(Electronic Edition), 2025, 19(02): 129-139.
[11]	Chunfeng Liu, Zhaohui Xu, Hongwei Shi, Rong Chen, Tengfei Ma, Pengfei Li, Rong Yuan, Jianrong Chen, Aiming Xu. Diagnosis of sarcopenia in mechanically ventilated patients and its impact on prognosis [J]. Chinese Journal of Clinicians(Electronic Edition), 2024, 18(09): 820-825.
[12]	Mingyuan Sun, Heng Chu, Haibin Xu, Zhe Zhang. Progress in application of artificial intelligence in diagnosis of multiple pulmonary nodules [J]. Chinese Journal of Clinicians(Electronic Edition), 2024, 18(08): 785-790.
[13]	Yiming Xu, Xueyuan Fan, Shihao Li, Yifan Xu, Bowen Li, Zhenyang Geng, Yafei Liu, Guanchao Ye, Feng Li, Lan Huang, Yu Qi. Artificial intelligence-driven neoadjuvant precision treatment of esophageal cancer [J]. Chinese Journal of Thoracic Surgery(Electronic Edition), 2025, 12(02): 96-104.
[14]	Qingwen Liu, Yong Han, Lidan Chen, Zhe Deng. Impact of early mechanical ventilation on mortality rates in adult in-hospital cardiac arrest: a retrospective cohort study [J]. Chinese Journal of Hygiene Rescue(Electronic Edition), 2024, 10(04): 203-206.
[15]	Xiaopeng Liu, Congyan Liu, Ning Yang, Chen Cai, Xiaobing Li, Hongyu Wang, Sisen Zhang. Effect of “three-acupoint and five-needle” therapy combined with abdominal pressure lifting on lung rehabilitation in patients with mechanical ventilation [J]. Chinese Journal of Hygiene Rescue(Electronic Edition), 2024, 10(04): 193-198.

Please choose a citation manager

Content to export