ICU细菌感染性肺炎病人抗菌药物相关性腹泻预警模型的构建

Construction of a warning model for antibiotic-associated diarrhea in ICU patients with bacterial pneumonia

目的:构建重症监护病房(ICU)细菌感染性肺炎病人抗菌药物相关性腹泻(AAD)的个体化预警模型。方法:选取赣州市人民医院ICU 2017年1月—2020年12月收治的317例细菌感染性肺炎病人,观察AAD发生情况,采用Logistic回归分析ICU细菌感染性肺炎病人发生AAD的危险因素,以R软件整合独立危险因素构建Nomogram模型,并对模型的精准度进行验证。结果:317例细菌感染性肺炎病人中,发生AAD 91例(28.71%);Logistic回归分析显示,年龄(≥70岁)、肺炎严重指数评分(≥108分)、抗菌药物种类(≥3种)、抗菌药物使用时间(≥10 d)以及清蛋白(≤32 g/L)是病人发生AAD的独立危险因素(P<0.05)。Nomogram模型的C-index为0.817,95%CI[0.788,0.853],模型预测结果与实际观测结果接近。决策曲线结果显示,高风险阈值>0.18时,Nomogram模型在预测ICU细菌感染性肺炎病人发生AAD方面高于单个独立危险因素的额外临床净收益。结论:年龄(≥70岁)、肺炎严重指数评分(≥108分)、抗菌药物种类(≥3种)、抗菌药物使用时间(≥10 d)以及清蛋白(≤32 g/L)是ICU细菌感染性肺炎病人发生AAD的独立危险因素,Nomogram模型可整合AAD的风险因素并用于临床风险分层及制定护理策略,具有较好的临床可操作性。

Objective:To construct an individualized early warning model for antibiotic-associated diarrhea(AAD) in patients with bacterial pneumonia in the intensive care unit(ICU). Methods:A total of 317 patients with bacterial pneumonia who were admitted to the ICU of Ganzhou People′s Hospital from January 2017 to December 2020 were selected,and the occurrence of AAD was observed.Logistic regression was used to analyze the risk factors of AAD in ICU patients with bacterial pneumonia. The R software was used to integrate independent risk factors to build a Nomogram model and verify the accuracy of the model. Results:Among 317 patients with bacterial pneumonia,AAD occurred in 91 cases(28. 71%). Logistic regression analysis showed that age(≥70 years old),pneumonia severity index score(≥108 points),types of antibiotics(≥3 types),duration of antimicrobial use(≥10 d),and albumin(≤32 g/L)were independent risk factors for AAD in patients(P<0. 05). The concordance index(C-index)of the Nomogram model was 0. 817,and the95%CI was(0. 788,0. 853). The predicted results of the model were close to the actual observed results. Decision curve results showed that when the high-risk threshold was>0. 18,the Nomogram model had an additional net clinical benefit over a single independent risk factor in predicting AAD in ICU patients with bacterial pneumonia. Conclusions:Age(≥70 years old),pneumonia severity index score(≥108 points),types of antibiotics(≥3 types),duration of antimicrobial use(≥10 d),and albumin(≤32 g/L)were independent risk factors of AAD in ICU patients with bacterial pneumonia. The Nomogram model could integrate the risk factors of AAD and be applied in clinical risk stratification and formulating nursing strategies,which have good clinical operability.