基于列线图模型分析老年重症脓毒症患者谵妄发生的风险因素

Risk factors for delirium in elderly patients with severe sepsis based on nomogram model analysis

摘要:

目的 基于列线图模型分析老年重症脓毒症患者谵妄发生的风险因素。方法 回顾性分析2021年10月—2023年10月新疆维吾尔自治区人民医院收治的308例老年重症脓毒症患者的临床资料，通过重症监护意识模糊评估法（CAM-ICU）评估患者是否发生谵妄，根据评估结果将患者分为谵妄组（103例）与非谵妄组（205例）。收集两组患者性别、年龄、受教育程度、基础疾病（高血压、高脂血症、冠心病、高血糖）、感染程度（严重脓毒症、脓毒症休克）、并发症（低氧血症、低蛋白血症、高碳酸血症、急性肾损伤、急性呼吸窘迫综合征）、是否使用镇静剂、基本生化指标等信息，并比较两组患者的序贯器官衰竭评估（SOFA）评分、炎症因子水平［C反应蛋白（CRP）、白细胞介素-6（IL-6）、降钙素原（PCT）］、可溶性蛋白-100β（S-100β）和同型半胱氨酸（Hcy）水平。采用多因素一般Logistic回归模型分析筛选患者发生谵妄的风险因素，构建列线图预测模型，并评价模型的拟合效果。结果 谵妄组年龄、SOFA评分、白细胞计数、CRP、IL-6、PCT、S-100β、Hcy水平均高于非谵妄组（P <0.05），血红蛋白水平低于非谵妄组（P <0.05）；谵妄组中脓毒症休克、合并高血糖、并发低氧血症、低蛋白血症和高碳酸血症、使用镇静剂和机械通气的占比更高（P <0.05）。谵妄组与非谵妄组在性别构成、BIM、吸烟、酗酒、高血压、高脂血症、冠心病、急性肾损伤、急性呼吸窘迫综合征、机械通气、连续肾脏替代、淋巴细胞计数及肌酐水平方面比较，差异均无统计学意义（P >0.05）。多因素一般Logistic回归分析结果显示，年龄、脓毒症休克、高血糖、高碳酸血症、低蛋白血症、使用镇静剂、SOFA评分，以及IL-6、S-100β、Hcy水平均是影响重症老年脓毒症患者发生谵妄的重要因素（P <0.05）。校准曲线显示，此预测模型预测的校正曲线趋近于理想曲，曲线的平均绝对误差为0.029，敏感性为79.6%（95% CI：0.705，0.869），特异性为93.2%（95% CI：0.888，0.962），约登指数为0.718，曲线下面积为0.932（95% CI：0.904，0.960）。结论 重症老年脓毒症患者谵妄的发生与年龄、血糖、营养状况、疾病严重程度、治疗方式、炎症水平等多种因素有关，针对这些指标的评估有助于早期识别谵妄。

Abstract:

Objective To analyze the risk factors for delirium in elderly patients with severe sepsis using a nomogram model.Methods A retrospective analysis was conducted on clinical data from 308 elderly patients with severe sepsis treated in our hospital from October 2021 to October 2023. Delirium was assessed using the Confusion Assessment Method for the ICU (CAM-ICU), and patients were categorized into delirium (n = 103) and non-delirium (n = 205) groups based on the assessment results. Data collected included gender, age, education level, underlying diseases (hypertension, hyperlipidemia, coronary heart disease, hyperglycemia, hypoglycemia), severity of infection (severe sepsis, septic shock), complications (hypoxemia, hypoalbuminemia, acute kidney injury, acute respiratory distress syndrome), use of sedatives or mechanical ventilation, and basic biochemical indicators. The Sequential Organ Failure Assessment (SOFA) score and levels of inflammatory markers [C-reactive protein (CRP), interleukin-6 (IL-6), procalcitonin (PCT)], soluble protein 100β (S-100β), and homocysteine (Hcy) were compared between the two groups. Logistic regression analysis was used to identify risk factors for delirium, and a nomogram prediction model was developed to evaluate the model's fit.Results The delirium group was older, had higher SOFA scores, white blood cell counts, CRP, IL-6, PCT, S-100β, and Hcy levels (all P <0.05), and lower hemoglobin levels compared to the non-delirium group (P <0.05). Higher proportions of septic shock, hyperglycemia, hypoxemia, hypoalbuminemia, and hypercapnia, as well as the use of sedatives and mechanical ventilation, were observed in the delirium group (P <0.05). No significant differences were found in gender composition, BMI, smoking, alcohol use, hypertension, hyperlipidemia, coronary heart disease, acute kidney injury, acute respiratory distress syndrome, mechanical ventilation, continuous renal replacement therapy, lymphocyte counts, or creatinine levels (P >0.05). Logistic regression analysis identified age, septic shock, hyperglycemia, hypercapnia, hypoalbuminemia, use of sedatives, SOFA score, and levels of IL-6, S-100β, and Hcy as significant factors influencing the occurrence of delirium in elderly patients with severe sepsis (P <0.05). The calibration curve showed that the predictive model closely approximates the ideal curve, with a mean absolute error of 0.029, sensitivity of 79.6% (95% CI: 0.705, 0.869), specificity of 93.2% (95% CI: 0.888, 0.962), Youden's index of 0.718, and an AUC of 0.932 (95% CI: 0.904, 0.960).Conclusion The occurrence of delirium in elderly patients with severe sepsis is associated with multiple factors including age, blood sugar levels, nutritional status, severity of illness, treatment methods, and inflammation levels. Assessing these indicators can help in the early identification of delirium.