研究背景與目的：近年來，有不少研究指出阻塞性睡眠呼吸中止疾患（簡稱OSA）患者呈現出高比例的憂鬱情緒，但其盛行率結果分歧，回顧過往文獻也發現OSA與憂鬱情緒的關聯性之結果亦呈現分歧的結果。值得注意的是，OSA與憂鬱情緒之間存在著類似的症狀表現——白天嗜睡，OSA患者可能依據OSA伴隨的嗜睡症狀，回應憂鬱問卷或診斷中的問題，而使其受到誤診或是使盛行率受到高估。為探討此議題，本研究進行兩項研究，分別以橫斷及長期追蹤的資料，探討（一）OSA患者的憂鬱情緒是否聚焦於身體面向，且OSA與憂鬱情緒之間是否受到白天嗜睡中介，以及（二）OSA患者睡眠檢查長期追蹤資料中憂鬱情緒的改變，是否與白天嗜睡程度有關。 研究方法：研究一以台北醫學大學附設醫院睡眠中心資料庫中，2010至2015年到台北醫學大學附設醫院睡眠中心看診並進行PSG檢測時所收集的資料（AHI、醒覺指標、缺氧指標、平均血氧飽和度、最低血氧飽和度、BDI-IA、ESS）進行分析，在排除小於20歲與睡眠疾患共病之後，共有2140位OSA（364女、1776男）患者資料被納入分析。本研究針對OSA患者的憂鬱情緒（BDI-IA）分數進行探索性因素分析，並以此因素結構與Beck與Steer（1993）所得之憂鬱情緒面向結構進行模型競爭，以求更適配於OSA患者之因素結構。再者，本研究取具輕度以上憂鬱情緒的OSA患者資料，以線性迴歸分析了解OSA嚴重度（AHI）、嗜睡程度（ESS得分）與憂鬱情緒（BDI-IA、身體面向、認知面向）之間的關聯性，並以拔靴法進行中介模型的檢定。此外，欲探討OSA患者的憂鬱情緒可能聚焦於身體面向，本研究亦將OSA患者依症狀嚴重度分組，並比較不同症狀嚴重度OSA患者的認知面向與身體面向憂鬱分數。研究二則是以台北醫學大學附設醫院睡眠中心資料庫追蹤一年至兩年間的OSA患者的資料進行資料分析，由於女性人數過少，因此僅納入81位男性OSA患者資料進行統計分析。 研究結果：研究一探索性因素分析結果顯示，「體重減輕」與「煩躁易怒」在因素分析結果的負荷量相當低，無法歸類至其一面向，予以排除，且「不滿自我」、「社交退縮」、「優柔寡斷」三題（原認知面向題項）在OSA患者樣本中被歸類至身體面向。競爭模型結果顯示，本研究所得之因素結構AIC值較低，因此以本研究所得之因素結構進行後續分析中認知面向與身體面向憂鬱情緒之計分與分析。以具憂鬱情緒的OSA患者資料進行相關分析結果發現，AHI 、覺醒指標、 缺氧指標、平均血氧飽和度、最低血氧飽和度皆與ESS呈現顯著相關；且ESS與BDI、身體面向、認知面向憂鬱情緒皆呈現顯著正相關。值得注意的是，OSA症狀嚴重度AHI雖未與BDI呈現顯著關聯性，但與身體面向憂鬱情緒達顯著正相關，且OSA病理機轉——醒覺指標、缺氧指標、平均血氧飽和度亦與身體面向憂鬱情緒達顯著關聯性。拔靴法結果顯示，在控制年齡與BMI後，ESS僅中介於醒覺指標與身體面向憂鬱情緒間的關聯性。將男女性分組後，女性患者的ESS中介於AHI與身體面向憂鬱情緒、醒覺指標與身體面向憂鬱情緒、以及缺氧指標與身體面向憂鬱情緒之間的關聯性，但男性OSA患者則無此中介效果。將OSA患者依症狀嚴重度分組，針對不同症狀嚴重度OSA患者的認知面向與身體面向憂鬱分數進行重複量數檢定，發現不論AHI為輕、中、重度，身體面向憂鬱分數皆較認知面向來得高。研究二檢視OSA患者一年至兩年間追蹤的結果，發現AHI改變量與BMI改變量，無論對於ESS的改變量、BDI-IA總分的改變量、認知面向以及身體面向憂鬱情緒的改變量，皆無顯著相關性；而ESS改變量則與BDI-IA總分的改變量、認知面向憂鬱情緒的改變量、身體面向憂鬱情緒的改變量呈現顯著相關性。依ESS改變量分組後，進一步以單因子變異數分析對BDI-IA改變量、認知面向憂鬱情緒改變量、身體面向憂鬱情緒改變量進行檢定，結果顯示，無論在BDI-IA改變量、認知面向憂鬱情緒改變量、或身體面向憂鬱情緒改變量，皆達顯著差異；而事後檢定顯示，無論在BDI-IA改變量、認知面向、身體面向憂鬱情緒，ESS升高組與ESS無變化組、ESS降低組呈現顯著差異，而ESS無變化組則與ESS降低組無顯著差異。 結論：本研究OSA患者自評BDI-IA的結果發現有高達35%的患者至少有輕度以上的憂鬱情緒，且本研究發現在女性OSA患者身上，OSA嚴重度、睡眠片段化病理機轉，與憂鬱情緒間，受到白天嗜睡程度中介；且OSA患者的憂鬱情緒明顯聚焦於身體面向上，此結果說明OSA患者的憂鬱情緒可能受到白天嗜睡程度所影響，本研究在於提醒臨床工作者對OSA患者進行憂鬱症臨床診斷時，需注意與釐清OSA患者是否以白天嗜睡的情況，來回應憂鬱症狀相關的嗜睡表現，以降低誤判OSA患者憂鬱情緒的可能性。 / Aims: Recent years, researchers found high prevalence of depression occurred in OSA patients. However, the inconsistency was also found in depression prevalence and in the association between OSA and depressive mood. Notably, excessive daytime sleepiness is one of common symptoms of both OSA and depression. High prevalence of depression might be an overestimation due to excessive daytime sleepiness reported by OSA patients. Two studies (cross-sectional and longitudinal studies) were conducted to examine (1) whether the depressive symptoms in OSA patients are more somatic in nature and whether mediation effect between OSA and depression exist, and (2) whether the changes in depressive mood correlate with the changes in excessive daytime sleepiness showed in long term follow-up data. Methods: Sleep test data (including AHI, arousal index, desaturation index, mean SaO2, lowest SaO2, BDI-IA, ESS) retrieved from database of Taipei Medical University Hospital Sleep Center from year 2010 to 2015 was used in Study 1. 2140 OSA patients (364 F、1776 M) were included after rule out patients who is under 20 year-old or comorbid with other sleep disorders. Exploratory factor analysis was conducted to extract the dimensions of depressive mood in OSA patients and the dimensions obtained were further compared with those of Beck and Steer (1993) through competing models. Furthermore, correlations between OSA severity, OSA pathological mechanism, excessive daytime sleepiness, and depressive mood were analyzed in the data of OSA patients with depressive mood, and boostrapping method was conducted to test mediation effect. For examining whether the depressive symptoms in OSA patients are more somatic in nature, cognitive dimension and somatic dimension were compared within different OSA severity. One to two year follow-up sleep test data retrieved from database of Taipei Medical University Hospital Sleep Center was analyzed in Study 2. There were only 81 male OSA patient data included due to the small numbers of female patients. Results: Exploratory factor analysis in Study 1 showed that “weight loss” and “irritability” should be excluded due to low factor loading in depressive mood of OSA patients. “Dissatisfaction”, “social withdrawal” and “indecisiveness” were categorized into somatic dimension in OSA patients. Competing models indicated the factors obtained in exploratory factor analysis were preferred due to lower AIC value, which suggested the model was better fit to OSA patients. Correlation analysis showed that AHI, arousal index, desaturation index, mean SaO2, lowest SaO2, BDI-IA score, somatic dimension and cognitive dimension significantly correlated with ESS. Notably, although the correlation between AHI and BDI-IA was non-significant, AHI positively correlated with somatic dimension. OSA pathological mechanism (arousal index, desaturation index, mean SaO2) also significantly correlated with somatic dimension on BDI-IA. After controlling age and BMI, mediation effects of excessive daytime sleepiness were only found on the relation of arousal index and somatic dimension. Mediation effects were also found on the relation of AHI and somatic dimension, on the relation of desaturation index and somatic dimension, and on the relation of arousal index and somatic dimension in female OSA patients, but the mediation effect was not found in male OSA patients. Furthermore, 3 X 2 ANOVA repeated measurement showed somatic dimension was significantly higher than cognitive dimension on BDI-IA in different AHI severity group. Follow-up data in Study 2 showed the change of AHI and BMI were not correlated with the change of ESS, BDI-IA, cognitive dimension and somatic dimension on BDI-IA. However, the change of ESS was significantly correlated with the change of BDI-IA, cognitive dimension and somatic dimension on BDI-IA. Furthermore, the change of BDI-IA, cognitive dimension and somatic dimension on BDI-IA were tested in different ESS change group. Results showed that the change of BDI-IA, cognitive dimension and somatic dimension on BDI-IA were significant in different ESS change group. Post-hoc analysis indicated the change of BDI-IA, cognitive dimension and somatic dimension in ESS elevated group were significantly different from ESS maintained group and ESS descend group but the difference between ESS maintained group and ESS descend group was non-significant. Conclusion: 35% of OSA patients were classified as co-occurring depressive mood using BDI-IA. However, mediation effect of excessive daytime sleepiness was found in female patients on the relation of OSA severity and depressive mood, and also on the relation of sleep fragmentation and depressive mood. Furthermore, depressive symptoms in OSA patients are more somatic in nature. The results showed depressive mood in OSA patients was probably affected by the excessive daytime sleepiness. The findings suggested the need to clarify the effect of excessive daytime sleepiness to prevent the overestimation of depressive mood in OSA patients.