多面向睡眠問題與憂鬱風險關聯性之探討

The Risk of Developing Depression with Multidimensional Measures of Sleep Problems

黃怡禎(I-Chen Huang)

指導教授 ： 郭柏秀

國立臺灣大學/公共衛生學院/流行病學與預防醫學研究所/碩士(2017年)

https://doi.org/10.6342/NTU201702040

摘要

失眠是常見而且影響甚廣的精神健康議題。失眠症與憂鬱症兩者之間關係緊密，近期的薈萃分析顯示，原發性失眠是憂鬱症的重要危險因子，相對危險性相當於兩倍。睡眠問題有許多面向，與憂鬱症的相關性可能因失眠症狀，睡眠連續性和客觀睡眠結構而有所不同。然而，很少有前瞻性研究以符合DSM-IV診斷標準的原發性失眠症患者做為研究族群，過去也沒有研究針對失眠症患者，發展個人層次的憂鬱風險預測模型。因此本研究透過前瞻性研究設計，旨於探討不同面向的睡眠問題與憂鬱症之間的關聯性，並且整合不同面向睡眠問題的訊息，建立罹患憂鬱症的風險預測模型。
本研究從精神科門診共招募56名符合DSM-IV診斷的原發性失眠症患者。在基線時所有患者都接受一夜多項睡眠監測檢查，以收集客觀睡眠連續性及睡眠結構參數，失眠症狀嚴重程度及主觀睡眠連續性參數則分別由中文版雅典失眠量表和匹茲堡睡眠品質量表進行評估，壓力則使用自覺壓力感知量表測量。追蹤訪查則以貝克憂鬱量表第二版評估憂鬱症狀的嚴重程度，並以探索式因素分析探討憂鬱症狀之因素結構。線性迴歸分析用於評估不同面向睡眠問題與憂鬱症之間的相關性，校正年齡、性別及其他潛在干擾因子，進一步年齡進行分層分析。最後依據基線的憂鬱程度分為基線有憂鬱組及基線無憂鬱組，利用逐步迴歸分析建立憂鬱風險預測模型，以接收操作特徵曲線下面積（Area under the receiver operating characteristic curve, AUC）檢測模型預測能力。
本研究發現基線的憂鬱程度（p=0.01）、焦慮程度（p<0.01）及壓力程度（p<0.01）可顯著預測未來的憂鬱風險。在睡眠相關變項發現，基線的日間功能損害（p=0.01）及主觀睡眠時數（p<0.05）可顯著預測未來的憂鬱風險，雖然在校正其他干擾因子後呈現不顯著結果。透過因素分析取得貝克憂鬱量表第二版共兩個因素，分別命名為「自我否定」與「失去能量」。基線的非快速動眼期第二期潛伏時間可顯著預測憂鬱之「自我否定」構面（p<0.05）。整合不同面向的睡眠問題建立憂鬱風險預測模型，基線無憂鬱組之模型包括變項有日間身心功能損害、非快速動眼期第二期潛伏時間、快速動眼期百分比及客觀睡眠時數，其AUC值為0.90，具有良好預測憂鬱風險之能力。
總結來說，基線的主觀及客觀的睡眠問題測量值皆與一年後的憂鬱風險增加有顯著相關性。期盼透過預測模型及早鑑別出高風險的原發性失眠症患者，提供更為貼切的治療策略，進而預防或減緩日後情感性疾病的發作。

關鍵字

失眠症 ； 憂鬱症 ； 預測 ； 前瞻性研究

並列摘要

Insomnia has been linked with an increased risk for depression in clinical and population samples, such that a two-fold relative risk of incident depression was reported in a meta-analysis of longitudinal studies among individuals with insomnia. There are many dimensions in sleep problems, and the associations with depression might vary for insomnia symptoms, sleep continuity, and objective sleep measures. Moreover, few prospective studies focus on patients with insomnia diagnosis based on DSM-IV criteria. So far, the individual-level predictive model for depression risk among insomnia patients is lacking in the literature.
The aims of the current study were to explore the relationships between different dimensions of sleep problems and depression in a prospective study design. In addition, we evaluated prediction models for the risk of depression using different dimensions of sleep measurements in patients with insomnia.
We recruited 56 patients of primary insomnia from psychiatry clinics. At baseline, all patients conducted an over-night polysomnography (PSG) examination to collect data on objective sleep continuity and sleep architecture parameters. Insomnia symptoms severity and subjective sleep continuity parameters were assessed using the Athens Insomnia Scale-Chinese version (CAIS) and Pittsburgh Sleep Quality Index (PSQI), respectively. Stress was measured using perceive stress scale. At follow-up, depressive symptoms severity was measured using Beck Depression Inventory-second edition (BDI-II). The factorial structure of the BDI-II was tested through exploratory factor analyses. Linear regression models were used to evaluate the associations between different dimensions of sleep problems and depression after adjusted for age, gender and potential confounders (e.g. stress and insomnia improvement). A stratified analysis was also performed by age. The prediction model was established by stepwise regression analysis and stratified by depression at baseline. Predictive ability was assessed by the area-under-the-receiver operating characteristic curve (AUC).
We found that depression score at baseline significantly predict depression at follow-up (p=0.01). In addition, higher perceived stress and anxiety at baseline also increased the risk of depression at follow-up (p<0.01). Among sleep variables, we found that daytime dysfunction (p=0.01) and subjective sleep duration (p<0.05) at baseline were significantly associated with elevated risk of depression at follow-up, though these associations turned insignificant with covariates adjustment. In younger age group, daytime dysfunction were significantly associated with elevated risk of depression at follow-up (p<0.05). Two factors were obtained for BDI-II by factor analysis, namely “self-negation” and “loss of energy”. We found that latency to non-rapid eye movement (NREM) stage 2 at baseline were significantly associated with elevated risk of “self-negation” at follow-up (p<0.05).
A prediction models was obtained for depression risk with good predictive ability. The AUC was 0.90 for prediction model among patients in the baseline non-depressed group. The model included daytime dysfunction, latency to NREM stage 2, rapid eye movement (REM) percentage and objective sleep duration.
In conclusion, both subjective and objective sleep measures at baseline are associated with an increased risk of depression after on average one-year of follow-up. We anticipate that the prediction models can assist to early identify primary insomnia patients with high risk of depression, and to provide appropriate treatment strategies to prevent the development of depression.

並列關鍵字

insomnia ； depression ； prediction ； prospective study

參考文獻

American Academy of Sleep Medicine. (2005). International Classification of Sleep Disorders, second edition: Diagnostic Coding Manual. Westchester IL.

American Academy of Sleep Medicine. (2007). The AASM Manual for the Scoring of Sleep and Associated Events- Rules, Terminology and Technical Specifications: AASM.

American Psychiatric Association. (2000). The Diagnostic and Ststistical Manual of Mental Disorder, Fourth Edition, Text Revision.

Aritake, S., Asaoka, S., Kagimura, T., Shimura, A., Futenma, K., Komada, Y., Inoue, Y. (2015). Internet-based survey of factors associated with subjective feeling of insomnia, depression, and low health-related quality of life among Japanese adults with sleep difficulty. Int J Behav Med, 22(2), 233-238. doi: 10.1007/s12529-014-9421-7

Bower, B., Bylsma, L. M., Morris, B. H., Rottenberg, J. (2010). Poor reported sleep quality predicts low positive affect in daily life among healthy and mood-disordered persons. J Sleep Res, 19(2), 323-332. doi: 10.1111/j.1365-2869.2009.00816.x