Data from: Age, gender, neck circumference, and Epworth sleepiness scale do not predict obstructive sleep apnea (OSA) in moderate to severe chronic obstructive pulmonary disease (COPD): the challenge to predict OSA in advanced COPD

The combination of chronic obstructive pulmonary disease (COPD) and obstructive sleep apnea (OSA) is associated with substantial morbidity and mortality. We hypothesized that predictors of OSA among patients with COPD may be distinct from OSA in the general population. Therefore, we investigated associations between traditional OSA risk factors (e.g. age), and sleep questionnaires [e.g. Epworth Sleepiness Scale] in 44 patients with advanced COPD. As a second aim we proposed a pilot, simplified screening test for OSA in patients with COPD. In a prospective, observational study of patients enrolled in the UCSD Pulmonary Rehabilitation Program we collected baseline characteristics, cardiovascular events (e.g. atrial fibrillation), and sleep questionnaires [e.g. Pittsburgh Sleep Quality Index (PSQI)]. For the pilot questionnaire, a BMI ≥25 kg/m2 and the presence of cardiovascular disease were used to construct the pilot screening test. Male: 59%; OSA 66%. FEV1 (mean ± SD) = 41.0±18.2% pred., FEV1/FVC = 41.5±12.7%]. Male gender, older age, and large neck circumference were not associated with OSA. Also, Epworth Sleepiness Scale and the STOP-Bang questionnaire were not associated with OSA in univariate logistic regression. In contrast, BMI ≥25 kg/m2 (OR = 3.94, p = 0.04) and diagnosis of cardiovascular disease (OR = 5.06, p = 0.03) were significantly associated with OSA [area under curve (AUC) = 0.74]. The pilot COPD-OSA test (OR = 5.28, p = 0.05) and STOP-Bang questionnaire (OR = 5.13, p = 0.03) were both associated with OSA in Receiver Operating Characteristics (ROC) analysis. The COPD-OSA test had the best AUC (0.74), sensitivity (92%), and specificity (83%). A ten-fold cross-validation validated our results. We found that traditional OSA predictors (e.g. gender, Epworth score) did not perform well in patients with more advanced COPD. Our pilot test may be an easy to implement instrument to screen for OSA. However, a larger validation study is necessary before further clinical implementation is warranted.

慢性阻塞性肺疾病（chronic obstructive pulmonary disease, COPD）与阻塞性睡眠呼吸暂停（obstructive sleep apnea, OSA）共病时，与显著的发病风险及死亡风险密切相关。我们假设，慢性阻塞性肺疾病患者群体中阻塞性睡眠呼吸暂停的预测因子，可能与普通人群中阻塞性睡眠呼吸暂停的预测因子存在差异。因此，我们针对44例晚期慢性阻塞性肺疾病患者，探究了传统阻塞性睡眠呼吸暂停危险因素（如年龄）与睡眠问卷（如爱泼沃斯嗜睡量表（Epworth Sleepiness Scale））之间的关联。本研究的次要目标为开发一项针对慢性阻塞性肺疾病患者的简化阻塞性睡眠呼吸暂停初步筛查测试。本研究为一项前瞻性观察性研究，研究对象招募自加州大学圣地亚哥分校（UCSD）肺康复项目，我们收集了受试者的基线特征、心血管事件（如心房颤动）以及睡眠问卷数据（如匹兹堡睡眠质量指数（Pittsburgh Sleep Quality Index, PSQI））。在初步筛查测试的构建中，我们采用体重指数（BMI）≥25 kg/m²以及心血管疾病史作为筛查指标。本队列中男性占比59%，阻塞性睡眠呼吸暂停患病率为66%。受试者的第1秒用力呼气容积（FEV1）均值±标准差为41.0±18.2%预计值，FEV1/用力肺活量（FVC）比值为41.5±12.7%。单因素logistic回归分析显示，男性性别、年龄增长以及颈围增大与阻塞性睡眠呼吸暂停无显著关联；爱泼沃斯嗜睡量表得分与STOP-Bang问卷得分同样与阻塞性睡眠呼吸暂停无显著关联。与之相反，BMI≥25 kg/m²（比值比OR=3.94，P=0.04）以及心血管疾病诊断史（OR=5.06，P=0.03）与阻塞性睡眠呼吸暂停存在显著关联，其曲线下面积（area under curve, AUC）为0.74。受试者工作特征（Receiver Operating Characteristics, ROC）分析显示，本研究开发的慢性阻塞性肺疾病-阻塞性睡眠呼吸暂停初步筛查测试（OR=5.28，P=0.05）与STOP-Bang问卷（OR=5.13，P=0.03）均与阻塞性睡眠呼吸暂停存在显著关联。其中，慢性阻塞性肺疾病-阻塞性睡眠呼吸暂停测试的曲线下面积（0.74）、敏感性（92%）与特异性（83%）均为最优。通过十倍交叉验证验证了本研究结果的可靠性。我们发现，传统的阻塞性睡眠呼吸暂停预测因子（如性别、爱泼沃斯量表得分）在病情更严重的慢性阻塞性肺疾病患者群体中表现不佳。本研究开发的初步筛查测试可能是一种易于实施的阻塞性睡眠呼吸暂停筛查工具，但仍需开展更大规模的验证研究，方可支持其进一步的临床应用。