维持性血液透析患者合并睡眠障碍的风险列线图模型的构建与验证

吴洁; 陈厚良; 杨艳娜

doi:10.3969/j.issn.1671-2390.2024.06.005

维持性血液透析患者合并睡眠障碍的风险列线图模型的构建与验证

Construction and validation of risk nomogram model for sleep disorders in maintenance hemodialysis patients

摘要

摘要:
目的分析维持性血液透析（maintenance hemodialysis，MHD）患者睡眠障碍的危险因素，并构建风险预测模型。
方法选取2018年1月至2022年12月期间收治的198例MHD患者为研究对象，采用回顾性分析法，依照患者是否合并睡眠障碍将患者分为睡眠障碍组（n = 134）及非睡眠障碍组（n = 64），通过单因素及多因素Logistic回归分析筛选出MHD患者睡眠障碍的危险因素；利用R软件建立MHD患者合并睡眠障碍的列线图模型，采用绘制受试者工作特征（receiver operating characteristic，ROC）曲线及校准曲线检验模型区分度及校准度，通过Bootstrap进行模型内部验证。
结果 198例MHD患者中，匹兹堡睡眠质量指数（pittsburgh sleep quality index，PSQI）总分为（10.32 ± 1.34）分，其中PSQI总分≥7分有134例，睡眠障碍发生率为67.68%（134/198）；单因素及多因素Logistic回归分析显示，透析时间（OR = 1.573）、合并皮肤瘙痒（OR = 3.044）、合并不宁腿综合征（OR = 2.722）、血清甲状旁腺素（OR = 1.030）是MHD患者发生睡眠障碍的危险因素（P<0.05）；血红蛋白（OR = 0.893）及Kt/V（OR = 0.031）是MHD患者发生睡眠障碍的保护因素（P<0.05）；风险预测模型ROC曲线下面积为0.901（P<0.001，95%CI：0.860～0.943），cut-off值1.69，灵敏度85.8%、特异度76.6%，Bootstrap内部验证法C-index指数为0.92，校正曲线与理想曲线拟合良好。
结论透析时间、血红蛋白、合并皮肤瘙痒、合并不宁腿综合征、血清甲状旁腺素、Kt/V等6个因素构成的列线图对MHD患者睡眠障碍风险具有较好的预测效能，可借助该模型对MHD患者睡眠障碍发生情况及危险因素进行预测及鉴别，为临床制定预防措施提供科学依据。

Abstract:
Objective To explore the risk factors for sleep disorders in maintenance hemodialysis （MHD） patients and construct a risk prediction model.
Methods A retrospective study was conducted among 198 patients who underwent MHD between January 2018 and December 2022. Based upon whether or not there were complications of sleep disorders, they were assigned into two groups of sleep disorder （n = 134） and non-sleep disorder （n = 64）. Uni/multivariate Logistic regression analyses were performed for identifying the risk factors for sleep disorders in MHD patients. A nomogram model for sleep disorders in MHD patients was constructed with software R. The discrimination and calibration of this model were tested using the receiver operating characteristic （ROC） curve and calibration curve. Internal validation of this model was conducted byBootstrap.
Results The total PSQI score was （10.32 ± 1.34）. There were 134 patients with PSQI score ≥7 and the incidence of sleep disorders was 67.68%（134/198）. Uni/multivariate Logistic regression analyses revealed that dialysis duration （OR = 1.573）, skin pruritus （OR = 3.044）, restless leg syndrome （RSL）（OR = 2.722） and serum parathyroid hormone （PTH）（OR = 1.030） were risk factors for sleep disorders in MHD patients （P<0.05）. And hemoglobin （Hb, OR = 0.893） and Kt/V（OR = 0.031） were protective factors （P<0.05）. The area under the ROC curve of risk prediction model was 0.901（P<0.001, 95%CI: 0.860-0.943）. Its cut-off value, sensitivity and specificity were 1.69, 85.8% and 76.6% respectively. C-index index of Bootstrap internal validation method was 0.92 and calibration curve fit well with an ideal curve.
Conclusion The nomogram model constructed based upon dialysis duration, Hb, skin pruritus, RSL, serum PTH and Kt/V has an excellent performance in predicting sleep disorders in MHD patient s. This model may be employed for predicting and identifying the occurrence and risk factors of sleep disorders in MHD patients, thus providing scientific rationales for clin ical prevention.

HTML全文

参考文献(20)

施引文献

资源附件(0)