睡眠至關重要：前瞻性研究探討多面向睡眠表現型對不同心血管疾病類型之影響

Abstract

引言： 心血管疾病長期位居全球主要死因之首，許多研究致力於識別其風險因子，並將相關發現納入臨床介入指引。而美國心臟協會近年將睡眠納入維護心血管健康的主要因子之一，並依照年齡層提出建議的睡眠時長。然而，健康的睡眠型態本質上是多面向的，涵蓋多種睡眠表現型。現今大部分研究聚焦於單一睡眠表現型對整體心血管疾病風險的影響，且多數依賴問卷調查來獲取主觀的睡眠資訊。因此，本研究旨在全面性的探討主觀及客觀睡眠表現型與各種心血管疾病亞型的關聯，藉此識別與心血管疾病相關性最強的關鍵睡眠表現型，並了解哪些心血管疾病亞型廣泛受到睡眠型態的影響，亦探討需重點介入的易感性族群。 方法： 本研究透過英國人體生物資料庫，以問卷調查取得了447,183名參與者的主觀睡眠資料，並進行長期追蹤。部分參與者則進一步配戴腕動計，有5天以上客觀睡眠資料的參與者，使用R套件GGIR取得有意義的睡眠參數。心血管疾病亞型則根據住院之ICD-10診斷碼及初級照護的病歷紀錄，使用PheCODE系統進行亞型分類。統計分析則利用Cox比例風險模型評估睡眠表現型與心血管疾病風險的相關性，並進一步計算族群可歸因分率，以量化特定睡眠表現型對疾病風險的貢獻。最後，針對性別及發病年齡進行分層分析，以識別易感性族群。 結果： 在主觀睡眠表現型中，午睡、失眠及晨起困難被發現為顯著的風險因子，且在共11種心血管疾病亞型之多重假設檢定經 Bonferroni 校正後（p < 0.0045），廣泛影響近半數的亞型。其中，失眠（HR = 1.14-1.30，PAF = 3.35-8.39）及午睡（HR = 1.05-1.20，PAF = 2.1-10.36）展現了相對較強的風險效應。客觀睡眠指標與心血管疾病的相關性則較弱，僅睡眠規律指數對週邊血管疾病呈現顯著的保護作用（HR = 0.71，p = 0.0029）。我們也觀察到缺血性心臟病與缺血性中風為最廣泛受到不良睡眠表現型影響的兩大心血管疾病亞型。 除此之外，結合五項主觀睡眠表現型計算的健康睡眠分數對整體心血管疾病、缺血性心臟病及缺血性中風均呈現顯著的保護效果。以健康睡眠分數3分以下定義為整體睡眠狀況差的參與者增加了13%整體心血管疾病的風險、22%缺血性心臟病風險以及12%缺血性中風風險。觀察到較高的族群可歸因分率則進一步表明，同時考量多面向的睡眠型態相較於單一睡眠表現型能更全面的反映睡眠對心血管的影響。而分層分析結果則發現失眠與午睡的風險效應在女性及早發性心血管疾病的參與者中影響更為顯著。 結論： 本研究結果突顯了睡眠在心血管健康中的多面向影響及其重要性。關鍵的睡眠表現型，特別是失眠與日間午睡，在調整其他共變項後，依然廣泛且顯著地提升心血管疾病的發生，尤其是在女性及早發性的群體中更為明顯。而客觀量測的睡眠規律也與週邊血管疾病存在顯著關聯。強調了整合主觀及客觀睡眠特徵以加強我們對睡眠健康理解的重要性，並凸顯在心血管疾病預防中實施更精準的睡眠介入策略的必要性。

Background Cardiovascular disease (CVD) continues to be the leading global cause of mortality. Although sufficient sleep duration has been recognized as essential for cardiovascular health, sleep health is inherently multidimensional, encompassing a range of sleep phenotypes. To date, most research has examined single sleep traits in relation to total CVD risk, often relying on subjective self-reports. This study aimed to (1) comprehensively assess both subjective and objective sleep phenotypes in relation to various CVD subtypes, (2) identify the key sleep traits most strongly associated with CVD onset and determine the subtypes most influenced by sleep health, also (3) explore population subgroups that may be more susceptible to sleep-related cardiovascular risk. Materials and Methods A prospective cohort study of 447,183 UK Biobank participants was conducted with extensive person-years of follow-up. Subjective sleep phenotypes were derived from self-reported questionnaires, while objective measures were extracted from a subset with over 5 days of actigraphy data using the R package GGIR. CVD subtypes were classified using the PheCODE system, based on inpatient ICD-10 and primary care records. Cox proportional hazards models were used to evaluate the associations between sleep phenotypes and CVD risk. Additionally, population attributable fraction (PAF) quantified the contribution of specific sleep phenotypes to disease burden. Finally, stratified analyses by age and gender were conducted to identify vulnerable populations. Results Among subjective phenotypes, napping, insomnia, and difficulty waking in the morning emerged as significant predictors, broadly influencing nearly half of all CVD subtypes after Bonferroni correction for multiple testing across 11 subtypes (p < 0.0045), with insomnia (HR = 1.14-1.30, PAF = 3.35-8.39) and napping (HR = 1.05-1.20, PAF = 2.10-10.36) exhibited relatively stronger risk effects. Objectively measured sleep features demonstrated weaker associations, with a significant protective effect only observed for sleep regularity index in relation to peripheral vascular disease (HR = 0.71, p = 0.0029). Additionally, ischemic heart disease and ischemic stroke were identified as the two CVD subtypes most widely affected by adverse sleep profiles. Notably, a healthy sleep score combining 5 subjective sleep phenotypes showed significant protective effects against total CVD, ischemic heart disease, and ischemic stroke. Participants with poor overall sleep health defined as a healthy sleep score of 3 or lower exhibited a 13% increase in total CVD risk, 22% in ischemic heart disease risk, and 12% in ischemic stroke risk. The higher PAF further suggested that a multidimensional assessment of sleep health better captured the overall impact of sleep on CVD risk. The stratified analyses revealed that the effects of insomnia and napping on CVD subtypes were more pronounced in females and in participants with earlier disease onset age. Conclusion Sleep plays a substantial and multidimensional role in cardiovascular health. Key sleep phenotypes, particularly insomnia and daytime napping, contribute broadly and significantly to CVD onset after adjusting for other covariates, especially among women and participants with early disease onset age. Additionally, objectively measured sleep regularity showed a significant association with peripheral vascular disease. These insights show the importance of integrating both subjective and objective sleep characteristics to deepen our understanding of sleep health, and emphasize the need for more targeted sleep interventions in CVD prevention strategies.