背景：阻塞性睡眠呼吸暂停(Obstructive Sleep Apnea, OSA)是一种需终生治疗的常见慢性疾病，在打鼾者女性中，OSA的患病率高达15.2%。但OSA在女性中的漏诊率超过90%，需寻找筛查效果好、简便易用的筛查工具。但目前常用的筛查工具对门诊打鼾女性的筛查效果尚不清楚。尽管已有研究者开发了针对女性人群的筛查工具，但均尚未成熟，在门诊打鼾女性中的筛查效果仍需验证。

目的：在门诊打鼾女性人群中检验并比较女性OSAHS筛查模型、NoASA评分和STOP-BANG问卷的筛查效果。

方法：连续收集260例于2018年3月至2019年1月在北京两家三级甲等医院接受多导睡眠监测的门诊打鼾女性，收集其问卷资料和PSG监测结果。根据PSG监测结果分析三种筛查工具的筛查价值，并使用ROC曲线判断三种筛查工具的最佳切分值，使用ROC曲线面积和Z检验判断最适合门诊打鼾女性的筛查工具。

结果：对三种筛查工具的筛查效果分别进行分析。三种筛查工具的得分均与诊断结果呈正相关（P<0.05）。女性OSAHS筛查模型的得分为7（4,10），AHI≥5次/小时时，模型的ROC曲线下面积（AUC）为0.769 (95%CI：0.711,0.828)，对OSA具有中度预测价值。而以6分为切分值时有最大正确指数（0.437），AUC、灵敏度、特异度分别为0.719、75.3%、68.4%，优于以7分为切分值时的筛查效果。NoASA评分的得分为6（2,9）。AHI≥5次/小时时，NoASA评分的AUC为0.654(95%CI：0.588,0.722)，对OSA具有低度预测价值。而以5分为切分值时有最大正确指数（0.236），AUC、灵敏度、特异度分别为0.618、81.5%、42.1%，AUC和灵敏度高于以7分为切分值时的值。STOP-BANG问卷的得分为3（2,5）。AHI≥5次/小时，STOP-BANG问卷的AUC为0.782(95%CI：0.726,0.838)，对OSA具有中度预测价值。而以4分为切分值时有最大正确指数（0.439），AUC、灵敏度、特异度分别为0.720、68.5%、75.4%，优于以3分为切分值时的筛查效果。以AHI≥5次/小时为诊断标准，比较三种筛查工具的ROC曲线下面积，其中STOP-BANG与女性OSAHS筛查模型的ROC曲线下面积无差异（P>0.0167），而两者与NoASA评分的ROC曲线下面积均有差异（P<0.0167）。

结论：女性OSAHS筛查模型和STOP-BANG问卷均对OSA具有中度的预测价值，NoSAS评分的预测价值较低。调整切分值后，女性OSAHS筛查模型在以6分为切分值时的灵敏度较高，能够提高打鼾OSA女性的检出率，可有望向社区进一步推广。

background obstructive sleep apnea (osa) is a common chronic disease required lifelong treatment. among women snorers, the prevalence of osa is as high as 15.2%. however, the missed diagnosis rate of osa in women is more than 90%.so it is necessary to find screening tools with better screening value. among the questionnaires that are commonly used for obstructive sleep apnoea screening, it remains unclear which one is more effective in terms of ease of use and diagnosis of female snores with obstructive sleep apnoea risk.

ive to validate and compare the osahs screening model of female snorers, noasa score and stop-bang questionnaire among outpatient snoring women..

methods to recruite 260 outpatient snoring women who underwent a polysomnography (psg) from march 2018 to january 2019, in two tertiary first-class hospitals in beijing. three questionnaires were completed. the screening value of the three screening tools was analyzed which based on the severity of osa, determined by apnea-hypopnea index (ahi). the area under curve(auc) of the receiver operating characteristics of three questionnaires were calculated to judge the most suitable screening tools for outpatient snoring women.

results ahi>/=5 events /h, the scores of the three screening tools were related to the diagnosis (p<0.05). the score of the osahs screening model of female snorers was 7 (4, 10). the auc of the screening model was 0.769 (95% ci: 0.711, 0.828), which was moderate predictive value. the youden index of 6 score was a maximum (0.437), and the auc, sensitivity and specificity of 6 score were 0.719, 75.3%, 68.4%, respectively, higher than those of 7 score. noasa score was 6 (2, 9), and the auc of noasa score was 0.654 (95%ci: 0.588, 0.722), which was low predictive value. the youden index of 5 score was a maximum (0.236), and the auc, sensitivity and specificity of 5 score were 0.618, 81.5%, 42.1%. auc and sensitivity were higher than those of 7 score.the score of stop-bang questionnaire was 3 (2, 5). the auc of stop-bang questionnaire was 0.782 (95% ci: 0.726, 0.838), which had moderate predictive value for osa. the youden index of 4 score was a maximum (0,439).the auc, sensitivity and specificity of 4 score were 0.720、68.5%、75.4%, higher than that of 3 score. the auc of three screening tools was compared. compared with the auc of the three screening tools, the auc of stop-bang and female osahs screening model had no difference (p>0.0167), but their auc differ from noasa score’s (p < 0.0167).

conclusion the osahs screening model of female snorers and stop-bang questionnaire have moderate predictive value for osa, while nosas score has low predictive value. after adjusting the cut-off value, the sensitivity of the osahs screening model of female snorers is higher when the cut-off value is 6.which can improve the detection rate of snoring osa women, and is expected to be further promoted to the community.