Sleep disorders and the risk of cognitive decline or dementia: an updated systematic review and meta-analysis of longitudinal studies

Abstract

Background

The evidence on the relationship between sleep disorders and the risk of cognitive decline or dementia remains inconsistent.

Objectives

This systematic review and meta-analysis aimed to provide updated evidence on the association between sleep disturbances and cognitive decline.

Methods

PubMed, EMBASE, and Web of Science were systematically searched from their respective inceptions to 18 February 2025. Cohort studies investigating longitudinal associations between sleep disorders and cognitive decline or dementia were included. Pooled relative risks (RRs) with 95% confidence intervals were calculated. Sensitivity analyses were conducted to evaluate the robustness of the pooled estimates. Publication bias was assessed using Egger’s and Begg’s tests, and meta-regression analysis was performed to explore potential sources of heterogeneity across studies.

Results

Seventy-six eligible cohort studies with eight types of sleep disturbances were included in the meta-analysis. Insomnia was associated with an increased risk of dementia (RR = 1.13). Both short sleep duration (7 h; RR = 1.27) and long sleep duration (8 h; RR = 1.23 for cognitive decline, RR = 1.43 for all-cause dementia, and RR = 1.66 for Alzheimer's disease (AD) were significant risk factors for cognitive decline and dementia. Excessive daytime sleepiness significantly increased the risks of vascular dementia (VD) (RR = 1.85), all-cause dementia (RR = 1.41), and cognitive decline (RR = 1.37). Sleep-related movement disorders indicated the strongest association, markedly increasing the risk of VD (RR = 2.53). Poor sleep quality was also a significant risk factor for AD (RR = 1.24), all-cause dementia (RR = 1.17), and cognitive decline (RR = 1.18).

Conclusion

This meta-analysis highlights sleep management as a pivotal modifiable factor in reducing the risk of all-cause cognitive decline. Systematic screening and early intervention for sleep disturbances should be prioritized as essential preventive strategies in clinical populations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00415-025-13372-x.

Introduction

Dementia is a progressive neurodegenerative syndrome characterized by persistent decline in cognitive function (memory and reasoning) and daily living capacities, ultimately impairing independent functioning [1]. According to the World Health Organization, over 55 million people worldwide are currently Living with dementia, with approximately 10 million new cases occurring annually [2]. The most common cause of dementia is Alzheimer's disease (AD), accounting for an estimated 60%–80% of cases, followed by vascular dementia (VD) and other subtypes [3]. Epidemiological projections indicate a staggering increase in disease burden, with many affected individuals expected to triple to 153 million by 2050 [4]. Dementia poses a significant threat to individuals and families, societies, and healthcare systems [5]. At present, there is no effective cure for dementia [6, 7]. Consequently, identifying modifiable risk factors is of paramount importance for delaying or preventing their onset [8].

Sleep disturbance is a potential modifiable risk factor for dementia [9]. Recently, numerous studies have investigated the relationship between sleep disorders, cognitive deterioration, and the risk of developing dementia [10–12]. These studies have revealed that sleep plays a vital role not only in physiological repair but also in maintaining cognitive health, through mechanisms such as the removal of metabolic wastes from the brain and the regulation of neuroplasticity [13, 14]. For example, sleep deprivation and circadian disruption have been found to decrease lymphatic clearance of harmful macromolecules such as β-amyloid and tau, increase oxidative stress in the brain, and decrease melatonin levels, which are necessary for circadian regulation [15]. These changes lead to neurodegenerative mechanisms and may increase the risk of dementia [12]. Therefore, identifying sleep-related risk factors may help identify individuals at increased risk for dementia and develop targeted prevention strategies.

Chronic sleep disorders impair cognitive function, accelerate cognitive decline, and promote the pathological changes characteristic of dementia. Notably, persistent sleep disorders often result in earlier institutionalization of individuals with AD, placing an enormous burden on families and society [16]. Sleep disorders cover a wide range of types, including insomnia, sleep-related breathing disorder (SRBD), excessive daytime sleepiness (EDS), sleep-related movement disorder (SRMD), and circadian rhythm sleep disorder, as classified in the International Classification of Sleep Disorders, third edition. Identifying which specific types of sleep disorders pose a higher risk of cognitive decline and dementia remains a critical issue requiring urgent attention and preventive measures.

Although numerous studies have investigated the association between sleep disorders and cognitive decline or dementia, several limitations have impeded the derivation of robust conclusions about the relationship between specific types of sleep disorders and the risk of cognitive decline or dementia. These limitations include insufficient sample sizes, variability in study designs (cross-sectional and cohort studies), heterogeneity in diagnostic criteria for sleep disorders, and inconsistent research findings. Consequently, a meta-analysis that systematically integrates and quantitatively synthesizes existing evidence is necessary to overcome these challenges. Such an approach holds significant theoretical and practical value for clinicians, enabling them to develop more precise diagnostic and treatment strategies.

While several systematic reviews have examined the association between sleep disorders and risk of cognitive decline or dementia [10, 11, 17, 18], important methodological limitations persist. First, many prior reviews focused on only a subtype of sleep disorders, failing to comprehensively address the broad spectrum of conditions classified under modern nosology. Second, variability in study designs (cross-sectional versus longitudinal approaches) and diagnostic methodologies has further complicated the comparability of findings. Moreover, over the past 5 years, numerous large-scale longitudinal studies have emerged, exploring the effect of various sleep-related exposures on incident cognitive disorders. Additionally, the results of the current review are contradictory. For example, one systematic review indicated that insomnia was associated only with incident AD [11], while the other systematic review revealed a significant association between insomnia and dementia risk, with increased risks for AD and VD [19]. Therefore, an update of the systematic review is necessary.

Herein, we conducted a meta-analysis to explore the risk of sleep disorders, including insomnia, SRBD, EDS, sleep quality, sleep duration, circadian rhythm, SRMD, REM sleep behavior disorder (RBD), and all-cause cognitive disorders based on longitudinal cohort studies.

Methods

Search strategy and selection criteria

We performed a systematic review and meta-analysis in accordance with preferred reporting items for systematic reviews and meta-analyses guidelines [20]. The complete PRISMA checklist is provided as supplementary material (Table S1). The protocol has been duly registered on the International prospective register of systematic reviews PROSPERO, with the registration number CRD42025631874.

A comprehensive search of PubMed, EMBASE, and Web of Science was performed through February 18, 2025, using the following strategy:

(“dementia” OR “Alzheimer’s disease” OR “cognition” OR “cognitive decline”) AND (“sleep disorders” OR “sleep disturbances” OR “sleep quality” OR “sleep duration” OR “insomnia” OR “sleep-disordered breathing” OR “SDB” OR “obstructive sleep apnea” OR “OSA” OR “central disorders of hypersomnolence” OR “circadian rhythm sleep disorder” OR “parasomnias” OR “sleep-related movement disorders” OR “sleep fragmentation” OR “excessive daytime sleepiness” OR “restless legs syndrome”) till 18 February 2025. Bibliographies of relevant original studies and systematic reviews were hand-searched in case of omission.

Additionally, related reviews and reference lists were checked manually.

Inclusion and exclusion criteria

Studies were included if they met all the following criteria: (1) Peer-reviewed English-language publications; (2) Population in this review were adults with diagnosed sleep disorders (e.g., insomnia, obstructive sleep apnea, sleep duration, EDS et al.) at baseline; (3) The study used a longitudinal design or prospective study that requires at least two measurements of cognition during follow-up; (4) The study explored the association of sleep conditions or parameters with risk of dementia or cognitive decline; (5) The outcomes of interest were risk of cognitive impairment and dementia (including Alzheimer's disease, vascular dementia, and other type dementia); (6) Reporting of multivariate-adjusted odds ratios (ORs), relative risks (RRs), or hazard ratios (HRs) with 95% confidence intervals (CIs).

Studies were excluded if they met any of the following criteria: (1) Risk estimate is not accessible; (2) Cross-sectional studies; (3) Only abstracts were available and (4) Case reports, commentaries, conference abstracts, and reviews; (5) Studies combining sleep disorders with comorbid conditions affecting cognition, and 6) studies for which the endpoint was not cognitive decline, all-cause dementia, AD, or VD.

Literature selection was performed by two experienced investigators (XYL and JO). If there were any discrepancies, consensus was reached through consultation with a third assessor (JHZ).

Data extraction

The following information was extracted for each study: (1) First author and publication year; (2) Study design and country); (3) Sample size (total/with sleep disturbances/dementia cases); (4) Participant demographics (sex ratio, age); (5) sleep disturbances assessed; (6) Clinical tools used for cognitive decline, dementia diagnosis; (7) Years of follow-up; (8) Covariates used for adjustment, and 9) Specific sleep disorder and cognitive outcome classifications.

Missing data were requested from corresponding authors. The data extraction was performed by two experienced investigators (XYL and JO) and any discrepancies were addressed by negotiation with the third reviewer (JHZ).

Assessment of study quality

Study quality was evaluated using the Newcastle–Ottawa Scale (NOS) by two researchers (XYL and JO) (NOS: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp) [21]. The NOS consists of nine items. Each item is assigned a star if the study meets the criteria of the item. The studies were divided into three categories: high quality (> 7 stars), medium quality (5–6 stars), low quality (< 4 stars) [21].

Definitions

We assessed the independent variables (i.e., sleep disorders, including insomnia, SRBD, EDS, sleep quality, sleep duration, circadian rhythm, SRMD, and RBD) with the dependent variables (dementia, including dementia of all causes, AD, and VD) and cognitive decline.

Insomnia refers to a clinical condition characterized by dissatisfaction with the quantity or quality of sleep, manifested as difficulty falling asleep, maintaining sleep, or waking up early and being unable to fall back asleep. SRBD is a group of disorders characterized by abnormal breathing during sleep, with OSA being the most typical and common. EDS refers to a condition where it is difficult to maintain wakefulness and alertness during primary waking periods. In situations where most people can easily stay awake, patients experience unwanted sleep episodes or fall asleep. Sleep quality refers to a multidimensional measure of subjective perceptions of sleep, including sleep depth, restorative quality (the extent to which one feels refreshed upon waking), and continuity, rather than merely sleep duration. It is a subjective concept. SRMD refers to a class of sleep disorders characterized by simple, repetitive movements that interfere with sleep initiation or maintenance. RBD refers to a parasomnia that occurs during REM sleep, characterized by dream-related behaviors (such as shouting, punching, kicking, or falling) and the absence of REM sleep-related muscle relaxation. The above diagnosis is mainly based on clinical scales and objective sleep monitoring, etc.

The diagnosis of dementia was required to be based on recognized diagnostic criteria, such as from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DMS-IV) criteria and the International Classification of Disease (ICD) codes. Participants were classified as having cognitive decline had a diagnosis of MCI or cognitive impairment. AD and VD, diagnosed primarily based on clinical scales and neuroimaging, are subtypes of dementia.

Statistical analysis

All statistical analyses were performed using Review Manager 5.4 software (The Cochrane Collaboration). For time-to-event outcomes, pooled estimates were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). To address discrepancies in reported effect measures across studies, we systematically converted odds ratios (ORs) to relative risks (RRs) using the following validated formula[22]:

$$\text{R}\text{Radjust}=\text{O}\text{Radjust}/\left[\left(1-\text{P}0\right)+\left(\text{P}0\ast \text{O}\text{Radjust}\right)\right].$$

where P0 represents the incidence of study endpoints (dementia or cognitive decline) in the non-exposed group. In cases where P0 values were unavailable, we utilized the overall sample incidence rate as a surrogate measure[22]. To enhance the clinical interpretability of our findings, 95% prediction intervals were calculated for all primary outcomes[23].

Heterogeneity was evaluated through Cochran's Q test with its corresponding p value and quantified using the I² statistic. The analytical approach was determined based on heterogeneity levels: a fixed-effects model was applied when homogeneity assumptions were met (I² < 50% accompanied by Q test p value ≥ 0.1). In cases of significant heterogeneity (I² ≥ 50% or Q test p value < 0.1), a DerSimonian–Laird random-effects model was adopted to account for between-study variance. This dual-method approach ensures appropriate model selection while addressing potential variability in effect estimates across studies.

In addition, meta-regression based on population characteristics (age, gender, follow-up time, and sample size), region APOE4 status, depression, outcome measurement tools were conducted to determine the source of heterogeneity.

Sensitivity and subgroup analyses

The robustness of finding was evaluated through leave-one-out sensitivity analysis to assess the influence of individual studies on the pooled estimates. When sufficient studies were available (n ≥ 10), meta-regression was conducted to explore potential effect modifiers.

Given the pathophysiological heterogeneity among sleep disorder subtypes and their differential associations with cognitive outcomes, we performed disease-specific subgroup analyses (e.g., insomnia vs. sleep-disordered breathing vs. circadian rhythm disorders et al.).

Bias assessment

Publication bias was evaluated using a two-stage approach for meta-analyses containing ≥ 10 studies. First, funnel plot asymmetry was assessed using Egger's regression method. Second, the trim-and-fill method was applied to estimate the number of potentially missing studies, accompanied by contour-enhanced funnel plots to help distinguish publication bias from other sources of asymmetry.

Results

Searching results

Figure 1 illustrates the flow diagrams of the study selection process. The search yielded 1,982 articles after deduplication. After scanning the titles and abstracts, 76 articles [24–99] were considered as potentially eligible. After reviewing the full texts, we further excluded 14 pieces of Literature for various reasons. After further integrating with an additional 15 papers from the reference of published reviews, 76 studies that met the inclusion and exclusion criteria were included in this meta-analysis.

Fig. 1.

Flow diagram showing the study selection process

Characteristics of eligible studies

The detailed characteristics of the included studies in the meta-analysis are indicated in Table 1. Most studies reported the association of sleep with dementia (36%), AD (20%), VD (9%), cognitive decline (34%), and only a few involved dementia or cognitive decline (3%) (Fig. 2A). In these studies, eight types of sleep disorders were identified: insomnia, sleep duration, SRBD, SRMD, circadian rhythm, sleep quality, RBD, and EDS (Fig. 2B). Regarding quality assessment, the Newcastle–Ottawa Scale score of all studies included in the meta-analysis is greater than 7, which indicates high quality. The total score of each study is indicated in Table S3.

Table 1.

Characteristics of 76 studies included in the meta-analysis

Study	Country	Total subjects	Female (%)	Age at baseline(y)	Sleep variables	Endpoint	Outcome assessment	Follow–up duration (y)	Adjusted confounding factors
Morgan et al., 1994[71]	UK	84	NA	≥ 65	Self-reported insomnia	All-cause dementia	DSM-IIIR	4	Age, sex, and premorbid cognitive status
Foley et al.,1999[47]	US	2905	0, 0%	77.1 ± 4.2	Symptoms of sleep apnea	All-cause dementia	DSM-III-R	3	Age, BMI, marital status, coronary heart disease, cognitive impairment, COPD, diabetes
Quesnot_Alperovitch 1999[79]	France	1389	815, 58.7%	59–71	Self-reported snoring and EDS	Cognitive decline	MMSE	4	Age, gender, educational level, BMI, high blood pressure, and initial value of MMSE
Cricco et al., 2001[42]	USA	6444	4015, 62.3%	72	Self-reported symptoms of insomnia	Cognitive decline	The nine-item version of Pfeiffer’s Short Portable Mental Status Questionnaire (SPMSQ)	3	Baseline score on Short Portable Mental Status Questionnaire, age, race, education level, household income, marital status, physical function, history of vascular disease, smoking status, alcohol consumption, and use of prescription sleep medications at third annual follow-up (FU3)
Foley et al., 2001[46]	US	2346	0, 0%	76.6 ± 3.9	Self-reported insomnia and EDS	MCI and All-cause dementia	CASI score and DSM-IIIR	3	Age, education, apolipoprotein E4 status, CASI score, depression, hours of sleep, daytime napping, coronary heart disease, and history of stroke from the baseline examination
Tworoger et al., 2006[92]	USA	1844	1844, 100%	70–81	Self-reported sleep duration and frequency of difficulties sleeping	Cognitive decline	TICS test	1.8	Age, education, smoking status, physical activity, high blood pressure, living status, mental health index from the SF-36, tranquilizer use, and alcohol consumption
Lobo et al., 2008[64]	Spain	4803	2771, 57.7%	73.5 ± 9.8	Self-reported sleep problems	All-cause dementia and AD	DSM-IV	2	Age, sex, and educational level
Benito-León et al., 2009[26]	Spain	3286	1870, 56.9%	≥ 65	Self-reported sleep duration	All-cause dementia and AD	DSM-IV	3.2	Age in years, educational level, smoker, and drinker
Elwood et al., 2011[45]	UK	1986	0, 0%	55–69	Self-reported snoring, sleep apnea, restless legs, insomnia, and EDS	Vascular dementia	NINCDS criteria	10	Age, social class, smoking, alcohol intake, BMI, angina, ECG ischemia, chest pain, national adult, reading test, and psychological distress
Osorio et al., 2011[75]	US	346	222, 64.2%	75.9 ± 6.5	Self-reported insomnia	AD	DSM-IV and NINCDS-ADRDA criteria	7.7	Depressed mood, age, sex, MMSE score, and ApoE4
Tranah et al., 2011[89]	US	1282	1282, 100%	82.69 ± 3.30	Objective dysregulated circadian activity rhythms	MCI and All-cause dementia	Petersen Criteria and DSM-IV	4.9	Age, clinic site, race, education, depression, BMI, self-reported walking for exercise, number of IADL impairments, benzodiazepine use, antidepressant use, sleep medication use, alcohol use, caffeine intake, smoking, self-reported health status, hypertension, and history of medical conditions
Yaffe et al., 2011[97]	US	298	298, 100%	82.3 ± 3.2	Objective SDB	All-cause dementia and MCI	Mini-Mental State Examination (MMSE), Trails B and DSM-IV	4.7	Age, race, BMI, education level, smoking status, diabetes, hypertension, antidepressant, benzodiazepine, and nonbenzodiazepine anxiolytics
Boot et al., 2012[32]	USA	651	196, 30.1%	77	Probable RBD using the Mayo Sleep Questionnaire	MCI	A consensus diagnosis of normal cognition, or 1 of the neurodegenerative syndromes (e.g., MCI, dementia, PD) as defined by published criteria	3.8	Age, sex, education, and medical comorbidity
Chen et al., 2012[40]	Taiwan	34,158	19,038, 55.7%	≥ 50	ICD-9 diagnostic insomnia	All-cause dementia	ICD-9	3	Age, sex, hypertension, type 2 diabetes mellitus, hyperlipidemia, and stroke
Jaussent et al., 2012[52]	France	4894	2790, 57.0%	≥ 65	Self-reported insomnia, quality, and EDS	Cognitive decline and all-cause dementia	Mini-Mental Status Examination (MMSE) and DSM-IV	8	MMSE at baseline, study center, sex, age, education, depression, BMI, chronic disease, mobility, living alone, apolipoprotein E ε4, consumption of fish, fruit and vegetables, and prescribed sleep medication
Keage et al., 2012[53]	UK	2012	1066, 53%	75	Self-reported insomnia, snoring, EDS, and sleep duration	Cognitive decline	Mini-Mental State Examination(MMSE)	10	Sex, age at baseline, World Health Organisation BMI classification, education, and cognition
Potvin et al., 2012[78]	Canada	1664	1159, 69.7%	65–96	Self-reported insomnia, quality, and sleep duration	Cognitive decline	Mini-Mental State Examination	1	Age, education, Mini-Mental State Examination score at baseline, depressive episodes, anxiety, psychotropic drug use, cardiovascular conditions, and chronic diseases (arthritis or rheumatism, eye diseases, backache or spinal problems, digestive problems, thyroid disorders, other metabolic disorders, anemia, hypercholesterolemia, asthma or emphysema or chronic bronchial diseases, liver diseases, kidney or urinary problems, skin diseases, and migraine or frequent headaches)
Benito-León et al., 2013[27]	Spain	2715	1545, 56.9%	72.9 ± 6.1	Self-reported sleep duration	Cognitive decline	Mini-Mental State Examination	3.4 ± 0.5	Age, gender, geographical area, educational level, diabetes mellitus, chronic obstructive pulmonary disease, depressive symptoms or antidepressant use, and medications with central nervous system effects
Chang et al., 2013[38]	Taiwan	8484	3450, 40.7%	≥ 40	ICD-9 diagnostic sleep apnea	All-cause dementia, AD and vascular dementia	ICD-9	5	Age, sex, hypertension, hyperlipidemia, diabetes, stroke, urbanization level, monthly income
Lim et al., 2013[61]	US	737	562, 76.3%	81.6 ± 7.2	Objective sleep fragmentation	AD	NINCDS-ADRDA criteria	3.3	Age, sex, education, depressive symptoms, Parkinson disease, stroke, hypertension, diabetes, coronary artery disease, congestive heart failure, cancer, thyroid disease, antidepressant, sedative-hypnotic, and anxiolytic medications
Peters et al., 2013[77]	USA	230	115, 50%	81.9 ± 5.1	Self-reported sleep disturbance	All-cause dementia, AD and VAD	DSM-III-R, NINCDS–ADRDA and NINDS-AIREN	3.3	Age, education, APOE Status, and 3MS
Sterniczuk et al., 2013[85]	Europe	28,697	15,639, 54.5%	64.6 ± 9.9	Self-reported sleep disturbance	AD or dementia	question in SHARE’s wave 4 questionnaire	4.3	Age, sex, education, BMI, and cognition
Virta et al., 2013[93]	Finland	2336	1118 47.9%	52.3 ± 6.2	Self-reported poor sleep quality, snoring and sleep duration	AD	Prescribed cholinesterase inhibitors or memantine	22.1	Age, sex, education, ApoE status, follow-up, life satisfaction, obesity, hypertension, physical inactivity and drinking
Blackwell et al., 2014[29]	USA	2822	0, 0%	76.0 ± 5.3	Objective insomnia, sleep duration and quality measured by wrist actigraphy and self-reported EDS	Cognitive decline	The Trails B and the 3MS	3.4 ± 0.5	Age, clinic site, race (white versus nonwhite), body mass index, education, number of depressive symptoms, number of comorbidities, presence of impairment of instrumental activities of daily living, benzodiazepine use, antidepressant use, self-reported health status, physical activity, alcohol use, and smoking status
Hahn et al., 2014[51]	Sweden	214	172, 80.4%	83.4 ± 4.9	Self-reported sleep pattern change	All-cause dementia and AD	DSM-IIIR and NINCDS-ADRDA criteria	9	Age, gender, education, physical function, pain, and breathing problems
Benedict et al., 2015[25]	Sweden	1574	0, 0%	49.6 ± 0.6	Self-reported insomnia	All-cause dementia, AD and vascular dementia	DSM-IV and NINCDS-ADRDA criteria	40	Age, BMI, leisure-time physical activity, educational level, hypertension, and diabetes
Blackwell et al., 2015[30]	USA	2636	0, 0%	Age at baseline(y)	SDB measured using in-home polysomnography	Cognitive Decline	Trails B and the 3MS	3.4 ± 0.5	Age, site, race, body mass index, education, number of depressive symptoms, history of diabetes mellitus, history of stroke or transient ischemic attack, history of hypertension, history of coronary heart disease, history of Parkinson’s disease, impairment in instrumental activities of daily living, benzodiazepine use, antidepressant use, self-reported health status, physical activity, alcohol use, and smoking status
Lin et al., 2015[62]	Taiwan	3020	1320, 43.7%	77.1 ± 4.2	ICD-9 diagnostic sleep related movement disorders	All-cause dementia, AD and vascular dementia	ICD-9	5	Age, gender, and comorbidities including hypertension, diabetes, ischemic heart diseases, hyperlipidemia, smoking, alcoholism, obesity, atrial fibrillation, Parkinson disease, cerebrovascular accident, major depression, chronic kidney disease
Martin et al., 2015[69]	France	559	337, 60%	59–71	SDB measured by an unattended ambulatory nocturnal respiratory recording	Cognitive impairment	Neuropsychological Assessment	7.8 ± 0.9	Sex, educational level, baseline age, number of years of follow-up and baseline clinical data (BMI, ESS, hypertension, diabetes, anxiety, and depression)
Song et al., 2015[84]	USA	2601	0, 0%	72	Sleep stages identified by in-home polysomnography	Cognitive decline	The Trails B and the 3MS	3.4 ± 0.5	Age, site, race (white versus nonwhite), education, depression, history of hypertension, history of angina, history of myocardial infarction, history of congestive heart failure, history of stroke/transient ischemic attack, benzodiazepine use, antidepressant use, self-reported health status, physical activity, and smoking status
Tsapanou et al., 2015[91]	US	1041	727, 69.8%	76.6 ± 3.9	Self-reported insomnia, restless legs and EDS	All-cause dementia	DSM-IIIR and NINCDS-ADRDA criteria	3	Age, gender, education, ethnicity and apolipoprotein E4 status, depression, stroke, hypertension, diabetes, and heart disease
Yaffe et al., 2015[98]	US	179,738	0, 0%	70–81	ICD-9 diagnostic sleep disturbance, sleep apnea and insomnia	All-cause dementia, AD and vascular dementia	ICD-9	8	Age, baseline diabetes, hypertension, myocardial infarction, cerebrovascular disease, obesity, depression, income tertile, and education
Chen et al., 2016[39]	USA	7444	7444, 100%	73.5 ± 9.8	Self-reported sleep duration	Cognitive decline (MCI or probable dementia)	Modified mini-mental state (3MS) examination, the validated 4-phase WHIMS protocols	7.3–7.7	Age, race, SES, lifestyle factors (smoking, alcohol consumption, and physical activities), depression, and other relevant clinical characteristics (previous HT use, BMI, prior CVD history, hypertension, DM, and hypercholesterolemia)
Diem et al., 2016[43]	USA	1245	1245, 100%	≥ 65	Sleep characteristics and sleep duration measured by actigraphy	MCI and dementia	MCI: a modified Petersen Criteria; Dementia: DSM-IV	4.9 ± 0.6	Age, race, clinic, education, body mass index, number of depressive symptoms, comorbidities, number of IADL impairments, smoking status, alcohol use, exercise, living alone, self-reported health status, antidepressant use, benzodiazepine use, and prescription sleep medication use
Ding et al., 2016[44]	USA, Canada, and Puerto Rico	7547	0, 0%	55–69	Self-reported SDB	Dementia	The Memory Impairment Screen (MIS)	5.7 ± 2.8	Age, race, BMI, education, APOE ɛ4 carrier, and self-reported indicators of cardio vascular disease (diabetes mellitus, hypertension, smoking)
Niu et al., 2016[73]	China	1010	656, 64.9%	75.9 ± 6.5	Self-reported sleep disturbance, quality and duration	Cognitive decline	Chinese version of the Mini-Mental State Examination (CMMSE)	1	Sex age, Mini-Mental State Examination (MMSE) score, education level, depression, snoring frequency, body mass index, physical activity, drinking status and smoking status, history of hyper tension, history of diabetes and history of coronary heart diseases, use of sleep medication
Bokenberger et al., 2017[31]	Sweden	11,247	6344, 56.4%	82.69 ± 3.30	Self-reported sleep quality and snoring	All-cause dementia	ICD-10	14.3	Age, follow-up time, sex, education, and baseline cognitive functioning
Gabelle et al., 2017[48]	France	479	325, 67.85%	82.3 ± 3.2	Self-reported EDS and RBD	Cognitive decline	Mini-Mental State Examination (MMSE)	1	Age, sex, educational level, CV and metabolic disorders and arm of randomization
Luojus et al., 2017[65]	Finland	2386	0, 0%	77	Self-reported sleep disturbance	All-cause dementia and AD	ICD-8, ICD-9 and ICD-10	21.9 ± 7.9	Age, examination year, Human Population Laboratory depression scale scores ≥ 5, physical activity, alcohol consumption, cumulative smoking history, systolic blood pressure, body mass index, low-density and high-density lipoprotein cholesterol, high-sensitivity C-reactive protein, cardiovascular disease history, education years and living alone
Pase et al., 2017[76]	USA	321	160, 50%	≥ 50	REM sleep measured by home-based PSG	All-cause dementia and AD	DSM-IV	12 ± 5	Age, sex, body mass index, education level, APOE ɛ4, smoking status, systolic blood pressure, treatment for hypertension, prevalent diabetes, prevalent heart disease, depressive symptoms, sleeping medication use, antidepressant use, and anxiolytic use
Reijs et al., 2017[80]	Europe	464	264, 57%	≥ 65	Self-reported sleep problem	AD	NINCDS-ADRDA	2.4 ± 1.3	Age, gender, education, center, and diagnosis
Westwood et al., 2017[94]	USA	2457	1400, 57%	75	Self-reported sleep duration	All-cause dementia and AD	DSM-IV	10	Age, sex, education, APOE ɛ4 allele status, and homocysteine
Sung et al., 2017[87]	Taiwan	184,158	119,169, 64.7%	65–96	ICD-9-CM diagnostic non-apnea sleep disorder (NSD)	All-cause dementia, AD and vascular dementia	ICD-9-CM	11	Index date, age, sex, urbanization level, and major comorbid conditions of stroke, diabetes, hypertension, hyperlipidemia, coronary artery disease (CAD), heart failure, atrial fibrillation, peripheral artery disease, obesity, mental illness and alcohol-attributed diseases (AADs)
Burke et al., 2018[35]	USA	12,083	7,865, 65.1%	72.9 ± 6.1	Self-reported sleep disturbance	Probable AD	NA	4.24	Age, race, education, ε4 carrier status, the presence of hypertension and hypercholesterolemia
Larsson_Wolk_2018[56]	Sweden	28,775	13,401, 46.6%	≥ 40	Self-reported sleep duration	AD	Record linkage with the Swedish National Patient Register	12.6	Age, sex, education, body mass index, and history of hypertension, hypercholesterolemia and diabetes, and mutually for the other lifestyle factors (except the Mediterranean diet score due to strong correlation with the DASH diet score) and sleep duration
Li et al., 2018[59]	USA	2461	1241, 50.4%	81.6 ± 7.2	Self-reported sleep duration	Dementia	NA	30	Age, gender, marital status, BMI, smoking, alcohol, low salt diet, and activity
Li et al., 2018[60]	USA	1097	844, 76.9%	81.9 ± 5.1	Sleep fragmentation index	MCI, AD	MCI: Diagnosis of MCI was rendered for persons who were judged to have cognitive impairment by the neuropsychologist but did not meet criteria for dementia by the clinician AD: NINCDS-ADRDA	11	Age, sex, education, and sleep fragmentation index
Lutsey et al., 2018[66]	US	1667	877, 52.6%	64.6 ± 9.9	Self-reported sleep duration and OSA confirmed by in-home polysomnography	MCI, All-cause dementia and AD	NIA-AA criteria	15	Age, sex, center, education attainment, the APOE ε4 risk allele, BMI, smoking status, leisure-time physical activity, diabetes, antihypertensive medications, C-reactive protein, and systolic blood pressure
Lysen et al., 2018[68]	Netherlands	4835	2791, 58%	52.3 ± 6.2	Self-reported sleep quality	All-cause dementia and AD	DSM-III-R and NINCDS–ADRDA	8.5	Age, sex, education, smoking, employment, coffee consumption, alcohol consumption, activities of daily living, cardiovascular risk factors, MMSE score, CES-D score, prevalent anxiety disorders, and snoring
Ohara et al., 2018[74]	Japan	1517	850, 56%	76.0 ± 5.3	Self-reported sleep duration	All-cause dementia, AD and vascular dementia	DSM-IIIR	8.8	Age, sex, education level, systolic blood pressure, antihypertensive agent, diabetes mellitus, hypercholesterolemia, body mass index, electrocardiographic abnormalities, history of stroke, smoking habits, alcohol intake, regular exercise and hypnotic use
Sindi et al., 2018[82]	H70 and KP: Sweden, CAIDE: Finland	H70: 437 KP: 306 CAIDE: 703	H70:264, 60.4% KP: 259, 84.6% CAIDE: 455, 64.7%	83.4 ± 4.9	Self-reported insomnia and sleep duration	All-cause dementia	CAIDE: DSM-IV, H70 and KP: DSM-III-R	H70: NA KP: 9 CAIDE: 21 ± 4.9	Sex, physical activity (yes/no), cohabitant status (alone or cohabiting), cardio/cerebrovascular conditions (stroke, myocardial infarction, atrial fibrillation, cardiovascular surgery, heart failure, or diabetes; yes/no), use of hypnotics (yes/no), hopelessness/depressive symptoms (yes/no), and apolipoprotein E (APOE) ε4 status
Suh et al., 2018[86]	South Korean	2893	1612, 55.7%	49.6 ± 0.6	Self-reported sleep disturbance, quality and duration	Cognitive decline (MCI and dementia)	Dementia: DSM-IV-TR; MCI: consensus criteria proposed by the International Working Group on MCI	4	Age, sex, education, presence of apolipoprotein E ε4 allele, Geriatric Depression Scale score, Cumulative Illness Rating Scale, socioeconomic status, employment status, presence of cohabitants, smoking, drinking, physical activity, REM sleep behavior disorder screening questionnaire score, and STOP-BANG score
Lee et al., 2019[57]	Korea	4362	1030, 23.6%	Age at baseline(y)	ICD-10 diagnostic SDB	AD	ICD-10	14	Sex, age, income level, CVD (coronary heart disease, cerebrovascular disease and heart failure), and peripheral arterial disease, hypertension, type 2 DM, depression, BMI, smoking status, physical activity, and alcohol drinking
Nakakubo et al.,2019[72]	Japan	2,096	1108, 52.9%	≥ 65	Self-reported sleep duration and EDS	Cognitive decline	The National Center for Geriatrics and Gerontology Functional Assessment Tool (NCGG‐FAT)	4	Age, sex, body mass index (BMI), education, medication, medical history (heart disease, respiratory disease, diabetes), taking sleeping pills or other medication to help sleep, current drinking habit, current smoking habit, physical activity, gait speed, Geriatric Depression Scale (GDS) score and Mini-Mental State Examination score
Bubu et al., 2020[33]	US and Canada?	1043	506, 48.5%	77.1 ± 4.2	Self-reported clinical diagnosis of OSA	AD	DSM-III	5.5 ± 1.7	Age, sex, body mass index (BMI), education, continuous positive airway pressure (CPAP) machine use, baseline biomarker data, hypertension, diabetes, history of cardiovascular disease (e.g., including ischemic heart disease, heart failure, and stroke/transient ischemic attack [TIA]), alcohol use, and history of traumatic brain injury
Lysen et al., 2020[67]	Netherlands	1322	699, 52.9%	59–71	Actigraphy-estimated sleep and 24-h activity rhythms	All-cause dementia and AD	DSM-III-R, NINCDS-ADRDA	11.2	Age, sex, education, paid employment, self-reported physical activity, habitual alcohol consumption, body mass index, positive history of cardiovascular disease (transient ischemic attack [TIA], stroke, heart disease), smoking status, presence of hypertension, and presence of diabetes mellitus
Smagula et al., 2020[83]	US	1951	1192, 61.1%	72	Self-reported EDS	All-cause dementia?	The Clinical Dementia Rating (CDR®) Dementia Staging Instrument	10	Cigarette smoking, alcohol consumption, exercise, social engagement, symptoms of depression, overall health, dependence in instrumental activities of daily living (IADL) impairment, cardiovascular disease, cerebrovascular disease, and diabetes
Tsai et al., 2020[90]	Taiwan	19,890	6781, 34%	76.6 ± 3.9	OSA	AD	ICD-9-CM	5.44 ± 2.96	Gender, age, urbanization level, income, arrhythmia, coronary artery disease (CAD), diabetes mellitus (DM), hypertension, stroke, heart failure, hyperlipidemia, head injury, depression, and anxiety
Beydoun et al., 2021[28]	USA	9518	5245, 55.1%	70–81	Self-reported insomnia	All-cause dementia	HRS-based dementia diagnosis	10	Socio-demographic, lifestyle characteristics (smoking status, frequency of alcohol consumption and frequency of moderate and vigorous exercise) and health characteristics (BMI, presence of cardiovascular risk factors, self-rated health and depressive symptoms)
Robbins et al., 2021[81]	USA	2812	1653, 60%	73.5 ± 9.8	Self-reported snoring, sleep quality and duration	Dementia	NA	5	Age, marital status, race, education, health conditions, and body weight
Agudelo et al., 2022[24]	USA and Canada	1391	623, 44.8%	≥ 65	Risk for SDB assessed using a modified STOP-BANG	MCI	MMSE and CDR	NA	Age and sex, BMI, APOE4 allele number, race, ethnicity, education, and marital status, hypertension, cardiovascular disease, stroke, alcohol abuse, and smoking
Cavaillès et al., 2022[36]	France	6851	4118, 60.1%	55–69	Self-reported insomnia, EDS, and sleep quality	All-cause dementia, AD, and vascular dementia	DSM-IV	12	Study center, sex, mobility, and presence of the APOE-ε4 allele, the level of education, diabetes mellitus, BMI, cardiovascular disease, depressive status
Choe et al., 2022[41]	USA	1058	471, 44.5%	75.9 ± 6.5	Self-reported sleep disorder (insomnia, OSA, and RLS)	MCI and AD	MCI: MMSE and CDR; AD: NINCDS-ADRDA	4.2	Age, sex, education, apolipoprotein E ε4 status, vascular risk score, body mass index, geriatric depression scale, and use of sleeping pill
Kim et al., 2023[55]	Korea	12,478	8131 65.2%	82.69 ± 3.30	RLS	All-cause dementia, AD and vascular dementia	ICD-10	12	Charlson Comorbidity Index (CCI), a history of schizophrenia, mood disorders (depression and bipolar disorder), anxiety disorders, Parkinson’s disease, iron deficiency anemia, and sleep disorders (insomnia, hypersomnia, sleep-related breathing disorder, narcolepsy, sleepwalking, sleep terror, and nightmare)
Wong et al., 2023[95]	USA	6284	3486, 55.5%	82.3 ± 3.2	Self-reported sleep problems	All-cause dementia	NHATS algorithm	10	Age, gender, race and ethnicity, the highest level of education, total household income, marital status, metropolitan residence, self-rated overall health condition, BMI, activities of daily living (ADLs), proxy respondent, major depressive disorder, generalized anxiety disorder, history of heart attack, history of hypertension, and history of diabetes
Bulycheva et al., 2024[34]	Japan	13,601	7109, 52.3%	77	Self-reported sleep duration	Dementia	Long-term care insurance database	8 ± 1.3	Age, sex, body mass index, marital status, education, occupation, smoking, alcohol drinking, total physical activity levels, insomnia, morning fatigue, use of sleeping pills, disease history (myocardial infarction, stroke, diabetes mellitus, depression), and sleep duration
Cavaillès et al., 2024[37]	France	6171	3902, 63.2%	≥ 50	Self-reported EDS	All-cause dementia, AD and vascular dementia	DSM-IV	12	Study center, gender, educational level, APOE ɛ4, impaired mobility, alcohol consumption, further adjusted for sleep medication use, depressive symptoms, number of insomnia complaints, or loud snoring
Guo et al., 2024[50]	UK	502,383	273,169, 54.4%	≥ 65	Self-reported insomnia, snoring, and daytime dozing	AD and vascular dementia	The lists of clinical codes used to define the clinical end points were developed and validated by the UK Biobank Outcome Adjudication Group in conjunction with clinical experts	13.08	Age, sex, ethnicity, and vascular risk factors
Khaing et al., 2024[54]	Australia	2187	1155, 52.8%	75	Self-reported sleep duration and EDS	Dementia	ICD-10	6	Age, sex, and education, smoking, alcohol consumption, hypertension, diabetes mellitus, hypercholesterolemia, cardiovascular disease, cerebrovascular disease, and depression and anxiety
Leng et al., 2024[58]	USA	526	305, 58%	65–96	Sleep duration and sleep fragmentation assessed by wrist actigraphy and self-reported sleep quality	Cognitive decline	Montreal Cognitive Assessment (MoCA)	30	Age, sex, race, education, BMI, depression, physical activity, hypertension, and diabetes
Liu et al., 2024[63]	UK	57,502	51,638, 89.8%	72.9 ± 6.1	Accelerometer-measured circadian rest-activity rhythm	All-cause dementia	ICD-10	6.86	Age, sex, education level, ethnicity, Townsend index of deprivation, recruitment center, season of accelerometer wear, smoking status, alcohol intake, healthy diet score, obesity, history of diabetes, hypertension, depression, and APOE ε4 carrier
Miyata et al., 2024[70]	Japan	41,731	22,513, 53.9%	≥ 40	Self-reported sleep duration	All-cause dementia	The criteria of the Long-term Care Insurance (LTCI) system	5	Age, sex, body mass index, smoking habits, alcohol intake, green tea intake, coffee intake, exercise habits, living arrangement, psychological stress, and history of diabetes and hypertension
Tan et al., 2024[88]	Sweden	22,078	13,648, 61.8%	81.6 ± 7.2	Self-reported insomnia and sleep duration	All-cause dementia and AD	ICD-9, ICD-10, and DSM-IV	19.2	Age, sex, education level, depression, social isolation, body mass index, level of physical activity, smoking status, alcohol consumption, hypertension, and diabetes
Xiong et al., 2024[96]	England	7223	3928 54.4%	81.9 ± 5.1	Self-reported sleep disturbance and duration	All-cause dementia and AD	All-cause dementia cases were classified by reports of a physician diagnosis of dementia or by a score of ≥ 3.4 on the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) given by family members or caregivers	8.0 ± 2.9	Age, sex, education, wealth, marriage, BMI, smoking, drinking, physical activity level, depressive symptoms, chronic disease and the competing risk of death with multivariate Fine and Gray Sub distribution Hazards
Zhang et al., 2024[99]	UK	155,828	77,649, 49.8%	64.6 ± 9.9	Self-reported poor sleep patterns	All-cause dementia, AD and vascular dementia	ICD-9, 10	12	age, gender, BMI, smoking status, drinking status, Townsend deprivation index, ethnicity, education level, Vitamin D supplements, diet score, family history of dementia, and self-reported comorbidities (diabetes, hypertension, cancer, etc.)
Gao et al., 2025[49]	UK	451,250	220,579, 48.9%	52.3 ± 6.2	Self-reported snoring	All-cause dementia, AD and vascular dementia	the UK Biobank-linked hospital inpatient records and death registries, which were documented based on the International Classification of Disease (ICD) codes	13.6	Age, sex, ethnicity, education, BMI, Townsend deprivation index (TDI) quintiles, smoking status, alcohol consumption, living alone, and histories of depression, diabetes, hypertension, ischaemic heart disease, and stroke

Fig. 2.

Numbers of studies in types of cognitive decline and type of sleep-related problems. A Types of cognitive decline; B Types of sleep- related problems. RBD, REM sleep behavior disorder; EDS, excessive daytime sleepiness; SRMD, sleep-related movement disorders; SRBD, sleep-disordered breathing disordered; CRSWDS, circadian rhythm sleep–wake disorders

Sleep disorders and cognitive disorders

When no classification of cognitive impairment was performed, we found that insomnia (RR = 1.09, 95% CI = 1.03–1.16, I² = 77.7%), SRBD (RR = 1.23, 95% CI = 1.14–1.33, I² = 80.8%), SRMD (RR = 1.99, 95% CI = 1.41–2.81, I² = 87.8%), EDS (RR = 1.38, 95% CI = 1.25–1.54, I² = 50%), quality (RR = 1.18, 95% CI = 1.12–1.26, I² = 80.9%), and sleep duration (RR = 1.27, 95% CI = 1.19–1.36, I² = 58.9%) were a major risk factor for cognitive disorders (Fig. 3).

Fig. 3.

Forest plot of the association between sleep disorders and the risk of cognitive disorders

When no classification of sleep disorders was performed, we found that sleep disorders were a major risk factor for AD (RR = 1.25, 95% CI = 1.15–1.35, I² = 84.3%),VD (RR = 1.32, 95% CI = 1.16–1.51, I² = 70.5%), dementia (RR = 1.18, 95% CI = 1.12–1.24, I² = 83.8%), and cognitive decline (RR = 1.19, 95% CI = 1.14–1.25, I² = 60.6%).

We further conducted a subgroup analysis of cognitive impairment due to high heterogeneity and to obtain accurate sleep types for cognitive decline and dementia risk.

Insomnia and cognitive disorders

Insomnia was a major risk factor for dementia (RR = 1.13, 95% CI = 1.04–1.23, I² = 79.70%) (Fig. 4).

Fig. 4.

Forest plot of the association between sleep characteristics and the risk of cognitive disorders. RBD, REM sleep behavior disorder; EDS, excessive daytime sleepiness; SRMD, sleep-related movement disorders; SRBD, sleep-disordered breathing disordered; CRSWDS, circadian rhythm sleep–wake disorders

Additionally, sensitivity analysis indicated that the exclusion of any individual study did not affect the overall effect size (Figure S2). Meta-regression analysis revealed that region (p = 0.00) and age stage (p = 0.00) are potential factors contributing to heterogeneity. Publication bias was detected using Begg’s test (Table S5). However, the trim-and-fill method indicated that the results remained consistent after imputing missing records (Figure S1).

Furthermore, we conducted a subgroup analysis of insomnia and found that only difficulty in sleep initiation was a major risk factor for cognitive disorders (RR = 1.11, 95% CI = 1.01–1.20, I² = 64.3%) (Fig. 5).

Fig. 5.

Forest plot of the association between insomnia subtype and the risk of cognitive disorders. DIS, difficulty in sleep initiation; DMS, difficulty in maintaining sleep; EMA, early morning awakening

Sleep duration and cognitive disorders

Sleep duration 7 h is a major risk factor for cognitive decline (RR = 1.27, 95% CI = 1.12–1.42, I² = 36.6%), sleep duration 8 h is a major risk factor for all-cause dementia (RR = 1.43, 95% CI = 1.21–1.69, I2 = 61.7%), AD (RR = 1.66, 95% CI = 1.44–1.91, I² = 0%), and cognitive decline (RR = 1.23, 95% CI = 1.12–1.36, I² = 12.1%; Fig. 6).

Fig. 6.

Forest plot of the association between sleep duration and the risk of cognitive disorders

Additionally, meta-regression analysis revealed that age stage (p = 0.00) is a potential factor contributing to heterogeneity. Publication bias was not detected using Begg’s test (Table S5). Further, the trim-and-fill method demonstrated that the results remained consistent after imputing missing records (Figure S1). Sensitivity analysis indicated that the removal of any included studies did not alter the results of the overall effect size (Figure S2).

EDS and cognitive disorders

EDS is a major risk factor for VD (RR = 1.85, 95% CI = 1.39–2.47, I² = 0%), all cause dementia (RR = 1.41, 95% CI = 1.19–1.67, I² = 56.9%), and cognitive decline (RR = 1.37, 95% CI = 1.15–1.64, I² = 44.5%).

Additionally, sensitivity analysis indicated that the removal of any included studies did not alter the results of the overall effect size (Figure S2). Meta-regression analysis revealed that follow-up (p = 0.023) is a potential factor contributing to heterogeneity. Publication bias was detected using Begg’s test (Table S5). However, the trim-and-fill method demonstrated that the results remained consistent after imputing missing records (Figure S1).

SRMDs and cognitive disorders

SRMDs are a major risk factor for VD (RR = 2.53, 95% CI = 1.30–4.93, I² = 76.1%; Fig. 4). Due to the small number of studies, meta-regression and sensitivity analyses were not implemented.

SRBD and cognitive disorders

Breathing disorders are a major risk factor for AD (RR = 1.39, 95% CI = 1.16–1.68, I² = 87%) and cognitive decline (RR = 1.22, 95% CI = 1.06–1.41, I² = 62.8%; Fig. 4).

Meta-regression analysis revealed that region (p = 0.032), sample size (p = 0.002), and detection method (p = 0.006) are potential factors contributing to heterogeneity. Publication bias was detected using Egger’s test (p = 0.001; Table S5). However, the trim-and-fill method demonstrated that the results remained consistent after imputing missing records (Figure S1). Sensitivity analysis indicated that the removal of any included studies did not alter the results of the overall effect size (Figure S2).

Sleep quality and cognitive disorders

Sleep quality is a major risk factor for AD (RR = 1.24, 95% CI = 1.08–1.42, I² = 80.4%), all cause dementia (RR = 1.17, 95% CI = 1.03–1.32, I² = 90.7%), and cognitive decline (RR = 1.18, 95% CI = 1.09–1.27, I² = 60.2%; Fig. 4).

Besides, meta-regression analysis revealed that detection method (p = 0.003) and APOE4 (p = 0.006) are potential factors contributing to heterogeneity. Publication bias was not detected using Begg’s test (Table S5). Moreover, the trim-and-fill method demonstrated that the results remained consistent after imputing missing records (Figure S1). Sensitivity analysis indicated that the removal of any included studies did not alter the results of the overall effect size (Figure S2).

Other sleep-related problems and cognitive disorders

No significant associations were found between circadian rhythm disturbances or RBD and the risk of dementia or cognitive decline (Fig. 4).

Subgroup analysis based on risk estimates for the effect of self-report and objective instrument

Subgroup analysis of self-reported scales and objective instrument to diagnose the types of sleep disorders were performed (Fig. 7) and found that objectively measured insomnia (RR = 1.26, 95% CI = 1.15–1.40, I²= 26.1%) was a statistically significant risk factor for cognitive disorders. Additionally, self-reported assessments of SRBD (RR = 1.21, 95% CI = 1.09–1.34, I² = 81.6%), EDS (RR = 1.38, 95% CI = 1.25–1.54, I² = 50.0%), poor sleep quality (RR = 1.20, 95% CI = 1.11–1.29, I² = 75.6%), short sleep (RR = 1.22, 95% CI = 1.10–1.35, I² = 62.3%), and longed sleep (RR = 1.44, 95% CI = 1.32–1.58, I² = 44.3%) were also significant contributors to the risk of cognitive disorders.

Fig. 7.

Forest plot of subgroup analysis based on risk estimates for the effect of self-report and objective instrument

Discussion

Summary of the main results

In this study, we found evidence supporting six types of sleep disorders (insomnia, SRBD, SRMD, EDS, sleep quality, and sleep duration) as predictors of a higher risk of cognitive disorders in non-demented adults.

The quantitative meta-analysis indicated that insomnia is a major risk factor for all-cause dementia; SRBDs are primary risk factors for developing mild cognitive impairment and AD; SRMDs are a major risk factor for vascular cognitive impairment; EDS is a key risk factor for VD, dementia, and cognitive decline; Sleep duration < 7 h primarily increases the risk of cognitive decline; while sleep duration > 8 h mainly elevates the risk of AD, dementia, and cognitive decline.

Biological mechanism between sleep disorders and cognitive disorders

The process by which sleep disorders lead to cognitive impairment is a long one, as we found in this study with follow-up periods ranging from 1 to 13 years. This process also has its biological mechanisms, which mainly include the following three aspects: (1) Accumulation of Aβ and Tau proteins: Chronic sleep disorders (such as insomnia and OSA) impair the function of the glymphatic system in cerebrospinal fluid, which is a key mechanism for clearing metabolic waste (including Aβ and Tau proteins) from the brain [100, 101]. This decline in clearance efficiency leads to the slow, gradual accumulation of toxic proteins in the brain over the years, eventually reaching a critical point that triggers neuronal damage and cognitive decline; (2) Synaptic plasticity and neuroinflammation: Chronic sleep deprivation disrupts neurogenesis in the hippocampus, impairs long-term potentiation (LTP), and leads to a persistent state of chronic neuroinflammation [102, 103]. These changes are also progressive, gradually eroding cognitive reserves and ultimately resulting in the collapse of brain network connectivity efficiency and cognitive resilience; (3) Vascular damage: Especially in obstructive sleep apnea (OSA), intermittent hypoxia and blood pressure fluctuations continuously cause microdamage to cerebral vascular endothelium, promoting atherosclerosis, slowly compromising cerebral blood flow, and contributing to vascular cognitive impairment [104].

Therefore, a shorter time interval (e.g., several months) may be insufficient to capture the significant clinical cognitive decline triggered by these underlying pathological changes. A sufficiently long lag time (e.g., 3 years, 5 years, or longer) is necessary and biologically meaningful. It reflects the nature of neurodegenerative disease—a process driven by the accumulation of time-dependent pathological burden.

Insomnia and cognitive disorders

This study identified insomnia as a significant sleep-related risk factor for all-cause cognitive disorders. This finding is partially consistent with a recent systematic review that revealed a significant association between insomnia and dementia risk, particularly increased risks for AD and VD [19]. This discrepancy may stem primarily from the relatively limited studies included in the recent systematic review. Additionally, meta-regression analysis revealed that region and age stage are potential factors contributing to heterogeneity. No publication bias was detected using Begg’s test.

Mechanistically, previous research has demonstrated that compared to patients with non-insomnia, individuals with insomnia exhibit more severe atrophy in the hippocampal cornu ammonis (CA) 2–4/dentate gyrus (DG) regions, which correlates with impairments in verbal information processing, verbal memory, language processing, and visual memory [105]. Furthermore, during normal sleep, the human brain uses the glymphatic system to clear amyloid-beta (Aβ) peptides accumulated during wakefulness [106]. Sleep disturbances may disrupt or impair this process, leading to cerebral Aβ accumulation, a hallmark pathological feature of AD pathogenesis [107].

Sleep duration and cognitive disorders

In this study, the classification of sleep duration was obtained by self-reporting the average sleep duration per week and was divided into three main categories: short sleep (< 7 h), ideal sleep (7–8 h), and long sleep (> 8 h) [96, 108]. We observed that longer sleep duration (> 8 h) was associated with an increased risk of all-cause dementia, AD, and cognitive decline compared with normal sleep duration. Additionally, we found that shorter sleep duration (< 7 h) was not associated with future risk of all-cause dementia and AD, which is consistent with the results of a previous systematic review[108]. The role of prolonged sleep duration in the development of dementia remains unclear and is closely associated with age. In the population of advanced elderly individuals (aged ≥ 70 years), those with longer sleep duration tend to have a higher risk of developing dementia [109].

Although several potential biological pathways have been identified, the mechanisms underlying the association between sleep duration and cognitive decline remain unclear. Tang et al. found that longer sleep duration was associated with higher plasma levels of Aβ40 and total tau, a lower Aβ42/Aβ40 ratio, and smaller gray matter volume [36]. Short-term sleep deprivation has been linked to enhanced hippocampal synaptic plasticity, leading to subsequent impairment of cognitive function.

EDS and cognitive disorders

We found that EDS was associated with a higher risk of developing VD, dementia, and cognitive decline. These findings are consistent with previous research, which demonstrated that individuals with EDS have a higher risk of all-cause dementia, although a non-significant association was observed with AD [110].

A recent study revealed that EDS is associated with higher cerebrospinal fluid levels of neuroinflammatory biomarkers (interleukin [IL]−6, and overall score) [111]. Prolonged sleep duration represents a preclinical marker driven by the APOE ε4 carrier gene, which is associated with poorer prognoses in elderly patients with cognitive impairment [112]. Additionally, studies have indicated an association between sleep duration and atrophy of hippocampal subregions [113]. Another study suggests that frequent EDS is independently associated with future vascular events and stroke, particularly in healthy community-dwelling elderly subjects [114]. Moreover, EDS often co-occurs with conditions such as depression, poor sleep habits, obesity, cardiovascular diseases, obstructive sleep apnea (OSA), and the use of sleeping pills. Intriguingly, each of these comorbidities is independently associated with an increased risk of cognitive impairment [97, 115].

SRMDs and cognitive disorders

SRMD, mainly referring to restless legs syndrome (RLS) in this study, is a major risk factor for vascular cognitive impairment. Postmortem and neuroimaging studies reported that patients with RLS exhibited silent and chronic microvascular disease and gliosis in the brain [116, 117]. Studies have found that RLS is also strongly associated with high blood pressure, heart disease, and stroke, suggesting that RLS may have vascular lesions [118, 119]. Neuroimaging and autopsy studies have found that patients with RLS exhibit chronic and asymptomatic cerebral microvascular injury and neurogliosis, all of which are risk factors for VD [116, 117].

SRBDs and cognitive disorders

SMBD, primarily referring to OSA, was identified as a major risk factor for all-cause dementia, AD, and cognitive impairment, which aligns with a previous systematic review [120]. Studies have indicated that intermittent hypoxia-induced ischemic brain injury, sleep fragmentation, along with systemic inflammation and cardiovascular instability, are the underlying factors accounting for the association between OSA and poor cognitive function [121]. Previous systematic reviews have demonstrated that the levels of blood total Aβ, Aβ40, Aβ42, and total tau are significantly increased in patients with OSA compared to healthy controls. This finding suggests a potential link between OSA and abnormal amyloid and tau protein metabolism, which are key pathological hallmarks associated with neurodegenerative diseases [122].

Sleep quality and cognitive disorders

In this study, we grouped eligible studies in this review measuring sleep using parameters such as quality, efficiency, and disturbance.

Sleep quality is a composite measure of multiple sleep-related issues, rather than a specific symptom of poor sleep. It encompasses sleep deprivation, sleep inefficiency, and daytime sleepiness. In this study, we found that sleep quality is the primary risk factor for all-cause dementia, AD, and cognitive impairment, which is consistent with previous research [10].

Sleep quality is primarily assessed using subjective measures, including the most commonly used instruments, the Pittsburgh Sleep Quality Index [123], the Neuropsychiatric Inventory, and the Insomnia Severity Index [124]. Sleep quality is considered to be an individual's self-satisfaction with various aspects of the sleep experience [125]. Poor sleep quality has been linked to amyloid deposition, which contributes to AD [126]. Sleep quality regulated by IL-18/IL-12 axis activation could be a potential mechanism underpinning neurocognitive changes [127]. There is a need for a more consistent consensus on defining sleep quality and for updating our understanding based on new evidence in the future.

Other sleep disorders and cognitive disorders

Many studies of cognitive impairment due to RBD and circadian rhythm are extremely small for a comprehensive evaluation, and further longitudinal studies are warranted in the future.

Circadian rhythms are approximately 24-h cycles of physiological processes in most organisms that are produced endogenously and can be regulated by external cues [128]. Circadian dysfunction may contribute to neurodegenerative disease by altering sleep duration, leading to decreased nighttime sleep consolidation and increased daytime napping. Moreover, sleep deprivation causes increased Aβ and tau pathology in mouse models [129, 130].

RBD is a special neurological state that accounts for 20%–25% of the nighttime sleep duration in healthy adults. It plays a crucial role in a wide range of functions, from basic physiological mechanisms to complex cognitive processes [130]. One recent study found that altered RBD sleep microstructure was associated with positron emission computed tomography (PET) detected Aβ deposition and greater neurodegeneration [130].

In the future, it is necessary to further explore the circadian rhythm, rapid eye movement, and the risks of cognitive impairment.

Publication bias and heterogeneity

We analyzed publication bias using funnel plots and found publication bias for multiple sleep disorders on dementia and cognitive decline, suggesting that some studies may be underrepresented. Furthermore, we did a trim-and-fill method and found that more literature is needed to reduce publication bias. This suggests a future focus on preprinted versions and statistically non-differentiated results in publishing articles to provide more reliable results for the association between sleep disorders and dementia or cognitive decline.

Regarding the heterogeneity observed in our analysis, it was particularly pronounced across different types of sleep disorders. This variability may be largely attributed to inconsistencies in diagnostic criteria—both for sleep disorders themselves and for cognitive impairment—which often rely on subjective assessment scales. To better understand the sources of this heterogeneity, we conducted detailed subgroup analyses and meta-regression examining potential moderators including geographic region, age group, sample size, follow-up duration, assessment method, and adjustment for APOE4 genotype. However, these factors only partially accounted for the observed heterogeneity, suggesting the influence of other unmeasured or unreported variables. Future studies should aim to adopt standardized diagnostic frameworks incorporating objective sleep measures (e.g., polysomnography or actigraphy) to minimize variability in sleep disorder definitions. Similarly, the use of objective biomarkers for diagnosing cognitive decline and dementia will be crucial to improve consistency across studies. In summary, while heterogeneity in this field arises from multiple methodological and clinical sources, employing rigorous longitudinal designs, adhering to strict diagnostic criteria, ensuring large sample sizes, and implementing follow-up periods exceeding 5 years are key strategies for enhancing the reliability and comparability of future findings.

Strengths and limitations of the study

This study has several significant advantages: (1) Only longitudinal cohort studies were included. (2) The sleep disorder spectrum and various dementia types were fully covered, and subgroup analyses were conducted for each area. (3) Our review incorporates a significantly broader range of the most recent literature (including studies up to February 28, 2025) compared to other recent systematic reviews. (4) Subgroup analysis and meta-regression were performed for each type of sleep disorders.

Several limitations should be addressed in this study. Initially, a cohort study design was chosen to enhance accuracy by enabling longitudinal assessment. However, only baseline data were used for defining sleep disorders, resulting in a limited selection of data that did not capture potential changes over time. This should be followed up for long-term sleep monitoring in the future, and averaging may be appropriate. Second, although we did meta-regression to adjust for various confounders in our analyses, some important factors, such as APOE genotype, psychiatric disorders, and medications, were not considered due to insufficient data provided in the original literature. Third, we conducted a subgroup analysis of self-reported scales and objective polysomnography to diagnose the types of sleep disorders and found that self-report was not statistically different for cognitive impairment and dementia, while the results of the objective scales indicated a statistically significant difference. In contrast, most of the included studies used self-reported scales, which may have some inaccuracy in the results. Studies that objectively assess sleep are needed to clarify the relationship between sleep disorders and cognitive impairment and dementia.

Conclusion

In summary, our findings indicate that various types of sleep-related exposures are associated with an increased risk of all-cause cognitive decline and dementia. Future investigations should aim to elucidate the biological mechanisms underlying these associations to inform the development of targeted therapeutic interventions. Besides, implementation studies are warranted to optimize the clinical translation of these research findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

ZJH and OJ designed the study. WTY, DW, DLY, and LXY collected the data, XJP and LYF helped with statistical analysis. ZJH write original draft, XJP, YB, and YHB revised the manuscript. All authors reviewed the paper and approved the final version.

Funding

This work was supported by the National Natural Science Foundation of China (82305388), Shenzhen Excellent Scientific and Technological Innovation Talent Training Program (RCBS20231211090814025), Sanming Project of Medicine in Shenzhen (SZZYSM202311002), Construction Project of Guangdong Famous Traditional Chinese Medicine Inheritance Studio (Document No. 108 [2023] issued by the Guangdong Traditional Chinese Medicine Administration), Program of the Guangdong Provincial Administration of Traditional Chinese Medicine (No. 20251316), Special project of the “Building Peaks and Creating Expertise” Action Plan of Guangzhou University of Chinese Medicine (GZY2025GB0210).

Data availability

All data analyzed during this study are included in supplementary materials.

Declarations

Conflicts of interest

The authors have no conflict of interest to report.