Page 85 - 《中国药房》2022年23期

利奈唑胺相关低钠血症危险因素分析及其列线图模型建立 Δ

秦 艳 ，叶岩荣 ，沈 赟，陈 喆，盛慧洁，李晓宇，吕迁洲（复旦大学附属中山医院药剂科，上海 200032）
＊
#
中图分类号 R978.1；R969.3 文献标志码 A 文章编号 1001-0408（2022）23-2891-05
DOI 10.6039/j.issn.1001-0408.2022.23.14

摘 要 目的 分析利奈唑胺相关低钠血症的危险因素，建立利奈唑胺致低钠血症的列线图模型。方法 回顾性收集复旦大学附
属中山医院2019年4月至2021年5月接受利奈唑胺治疗的142例患者资料，包括患者的人口统计学特征、实验室检查指标、合并
疾病和联合用药等情况。根据是否发生低钠血症分为低钠血症组和非低钠血症组，采用单因素分析比较两组间变量的差异，通过
多因素Logistic回归分析研究利奈唑胺相关低钠血症的独立危险因素；在此基础上建立列线图模型用于识别利奈唑胺致低钠血症
的发生风险，采用受试者工作特征（ROC）曲线和校准曲线、Hosmer-Lemeshow拟合优度检验评估模型的预测效能。结果 142例
患者中，有30例患者发生低钠血症，112例患者未发生低钠血症，不良反应发生率为21.1%。单因素分析结果显示，低钠血症组与
非低钠血症组患者的利奈唑胺谷浓度、基线血清钠、白细胞计数、总胆红素、白蛋白、丙氨酸转氨酶、天冬氨酸转氨酶、联合使用螺
内酯比较，差异有统计学意义（P＜0.05）。多因素Logistic回归分析结果显示，利奈唑胺谷浓度[OR＝1.128，95%CI（1.028，1.239）]、
基线血清钠[OR＝0.719，95%CI（0.604，0.857）]和总胆红素[OR＝1.007，95%CI（1.002，1.011）]是利奈唑胺相关低钠血症的独立危
险因素（P＜0.05）。列线图模型ROC曲线下面积（95%CI）为0.874（0.794，0.995），灵敏度和特异度分别为88.4%和76.7%；校准曲
线的平均绝对误差为0.017；Hosmer-Lemeshow拟合优度检验结果显示，风险预测值与实测值比较，差异无统计学意义（χ ＝4.941，
2
P＝0.064）。结论 利奈唑胺谷浓度、基线血清钠和总胆红素是利奈唑胺相关低钠血症的独立危险因素；在此基础上建立的列线图
模型对于利奈唑胺相关低钠血症的发生具有较好的预测价值。
关键词 利奈唑胺；低钠血症；危险因素；不良反应；列线图模型

Risk factor analysis of linezolid-induced hyponatremia and the establishment of its nomogram model
QIN Yan，YE Yanrong，SHEN Yun，CHEN Zhe，SHENG Huijie，LI Xiaoyu，LYU Qianzhou（Dept. of Pharmacy，
Zhongshan Hospital Affiliated to Fudan University， Shanghai 200032， China）

ABSTRACT OBJECTIVE To analyze the risk factors of linezolid-induced hyponatremia， and establish nomogram model of
linezolid-induced hyponatremia.METHODS The clinical information of 142 patients who received linezolid therapy were collected
from Zhongshan Hospital Affiliated to Fudan University from April 2019 to May 2021 including demographic characteristics，
laboratory index， concomitant disease and drug combination. They were divided into hyponatremia group and non-hyponatremia
group according to whether hyponatremia occurred； univariate analysis was used to compare the differences of variables between
the two groups； the independent risk factors for linezolid-induced hyponatremia were analyzed by multivariate Logistic regression.
The nomogram model was set up to identify the occurrence risk of linezolid-induced hyponatremia， receiver operating characteristic
（ROC） curve and calibration curve， Hosmer-Lemeshow goodness-of-fit test were used to evaluate the predictive effectiveness of the
model. RESULTS Of 142 patients， 30 patients suffered from hyponatremia， and 112 patients did not suffer， the incidence of
adverse drug reaction was 21.1%. Univariate analysis showed that there was statistical significance in trough concentration of
linezolid， baseline serum sodium， white blood cell count， total bilirubin， albumin， alanine aminotransferase， aspartate
aminotransferase， combined use of spironolactone between hyponatremia group and non-hyponatremia group （P＜0.05）.
Multivariate Logistic regression analysis showed that trough concentration of linezolid [OR＝1.128， 95%CI（1.028-1.239）]， baseline
serum sodium [OR＝0.719， 95%CI（0.604-0.857）] and total bilirubin [OR＝1.007， 95%CI（1.002， 1.011）] were independent risk
factors for linezolid-induced hyponatremia （P＜0.05）. The area under the ROC curve （95%CI） of the nomogram model was 0.874
（0.794-0.995）； the sensitivity and specificity were 88.4% and 76.7%. The average absolute error of calibration curve was 0.017.
The results of Hosmer-Lemeshow goodness-of-fit test showed that there was no statistically significant difference between the
predicted risk value and the measured value （χ ＝4.941，P＝0.064）. CONCLUSIONS The trough concentration of linezolid，
2
baseline serum sodium and total bilirubin are independent risk factors for linezolid-induced hyponatremia. The established
nomogram model shows well predictive performance to identify linezolid-induced hyponatremia.
KEYWORDS linezolid； hyponatremia； risk factor； adverse drug reaction； nomogram model

Δ 基金项目 复旦大学附属中山医院项目（No.2020ZSQN53）
利奈唑胺是一种噁唑烷酮类抗菌药物，临床应用广
＊第一作者 主管药师，硕士。研究方向：医院药学。电话：021- [1]
64041990。E-mail：celeryan@163.com 泛，常用于治疗多药耐药的革兰氏阳性菌感染 。该药
与其他抗菌药物不易发生交叉耐药或不易诱导细菌耐
# 通信作者 副主任药师。研究方向：医院药学、药事管理。电话：
[2]
021-64041990。E-mail：ye.yanrong@zs-hospital.sh.cn 药性的产生 ，静脉或口服生物利用度均较高。利奈唑

中国药房 2022年第33卷第23期 China Pharmacy 2022 Vol. 33 No. 23 · 2891 ·