联合肺功能指标分析老年阻塞性睡眠呼吸暂停低通气综合征患者认知损害的影响因素

Analysis of influencing factors for cognitive impairment in older-adult OSAHS patients based on combined pulmonary function indicators

中文关键词: 阻塞性睡眠呼吸暂停低通气综合征    认知损害    气流受限    LASSO回归    随机森林

英文关键词: OSAHScognitive impairmentairflow limitationLASSO regressionrandom forest

中文摘要:

      目的 探讨气流受限及夜间呼吸紊乱对老年阻塞性睡眠呼吸暂停低通气综合征(OSAHS)患者认知损害的影响。方法 采用横断面研究设计，纳入2022年7月至2024年6月于上海市普陀区利群医院就诊的年龄≥60岁受试者102例，其中OSAHS组74例，对照组28例。收集所有受试者的一般临床资料、肺功能参数、睡眠呼吸监测指标及蒙特利尔认知功能评估量表(MoCA)评分。分析MoCA评分与肺功能指标的相关性;运用LASSO回归与随机森林算法筛选预测变量，通过受试者操作特征(ROC)曲线评价模型预测效能;将2种方法共同筛选的核心变量纳入多因素Logistic回归分析。结果 校正混杂因素后，OSAHS组MoCA评分显著低于对照组，认知损害发生率高于对照组。无气流受限亚组与有气流受限亚组的认知损害发生率分别为57.89%和80.56%。MoCA评分与多项肺功能指标呈显著正相关，其中用力呼出25%肺活量时的瞬时呼气流量及呼气流量峰值(PEF)的相关性最强(P <0.001);第1秒用力呼气容积、用力肺活量(FVC)、用力呼出50%肺活量时的瞬时呼气流量、最大呼气中期流量和用力呼出75%肺活量时的瞬时呼气流量亦呈显著正相关(P <0.01)。LASSO回归与随机森林共同筛选出5个核心因子:体重指数(BMI)、FVC、PEF、血氧饱和度下降指数(ODI)及最长呼吸暂停时间。2种模型均表现良好，LASSO模型与随机森林模型的ROC曲线下面积分别为0.778和0.861。多因素Logistic回归分析显示，BMI(OR=1.423,95%CI:1.053～1.922)和ODI(OR=1.154,95%CI:1.024～1.300)为认知损害的独立危险因素，而PEF(OR=0.447,95%CI:0.234～0.854)则为其独立保护因素。结论 老年OSAHS患者认知损害发生风险显著升高。呼气峰值流量降低与夜间氧减事件频发可独立影响患者认知水平。

英文摘要:

      Objective To investigate the effects of airflow limitation and nocturnal respiratory disturbance on cognitive impairment in elderly patients with obstructive sleep apnea hypopnea syndrome(OSAHS).Methods A cross-sectional study was conducted.102 subjects aged 60 or above who visited Shanghai Putuo Liqun Hospital from July 2022 to June 2024 were included.Among them,74 were in OSAHS group and 28 were in control group.Their general clinical data,pulmonary function indicators,sleep respiratory monitoring indicators,and Montreal cognitive assessment(MoCA)scores were collected.The correlation between MoCA scores and pulmonary function indicators was analyzed.LASSO regression and random forest algorithm were used to screen predictive variables,respectively,and the predictive performance of the models was evaluated using receiver operating characteristic(ROC)curves.The core variables selected by both methods were incorporated into the multivariate Logistic regression analysis.Results After adjusting for confounding factors,the MoCA score in the OSAHS group was significantly lower than that in the control group,and the incidence of cognitive impairment was higher.The incidence of cognitive impairment was 57.89％in the non-airflow limitation subgroup and 80.56％in the airflow limitation subgroup.MoCA scores were significantly positively correlated with multiple pulmonary function indicators,among which forced expiratory flow at 25％of vital capacity and peak expiratory flow(PEF)showed the strongest correlations(P<0.001).Forced expiratory volume in one second,forced vital capacity(FVC),forced expiratory flow at 50％of FVC,maximal mid-expiratory flow and forced expiratory flow at 75％of FVC were also significantly positively correlated with MoCA scores(P<0.01).LASSO regression and random forest jointly screened out five core factors:body mass index(BMI),FVC,PEF,oxygen desaturation index(ODI),and longest apnea time(LAT).Both models performed well,with area under the ROC curve values of 0.778 for the LASSO model and 0.861 for the random forest model.Multivariate Logistic regression analysis showed that BMI(OR=1.423,95％CI:1.053-1.922)and ODI(OR=1.154,95％CI:1.024-1.300)were independent risk factors for cognitive impairment,while PEF(OR=0.447,95％CI:0.234-0.854)was an independent protective factor.Conclusion Elderly OSAHS patients have a significantly increased risk of cognitive impairment.The decrease in PEF and frequent occurrence of nocturnal hypoxemia events can independently affect patients' cognitive level.