Prevalence and clinical implications of Rapid eye movement sleep behavior disorder in Alzheimer’s disease and Lewy body dementia

Abstract

Rapid eye movement sleep behavior disorder (RBD) is increasingly recognized as a significant clinical feature in neurodegenerative diseases like Alzheimer’s disease (AD) and Lewy body dementia (LBD). This study investigates the prevalence and clinical implications of RBD among individuals diagnosed with AD and LBD, emphasizing its relationship with cognitive decline. A retrospective observational study analyzed data from 5,481 participants across four diagnostic groups: AD + LBD, AD, LBD, and cognitively normal controls. Multinomial logistic regression and multivariable analyses were performed to assess RBD associations with demographic, clinical, and neuropsychiatric factors. Limitations include reliance on self-reported questionnaires to identify RBD, which may lead to underdiagnosis and skewed prevalence estimates. RBD prevalence was significantly higher in the AD + LBD (37.40%) and LBD (69.30%) groups compared to AD (3.61%) and cognitively normal controls (2.31%). The odds ratio (OR) for RBD in the AD + LBD group was 12.39 compared to controls (p < 0.001). Other significant associations included male gender and higher neuropsychiatric symptom severity, while age and education demonstrated complex relationships with RBD prevalence. This study highlights RBD as a critical factor in neurodegenerative disease assessment and an early marker for identifying at-risk individuals. Recognizing RBD’s role in cognitive decline emphasizes the need for integrated clinical approaches that incorporate sleep assessments in dementia management and inform strategies for improving patient quality of life.

Introduction

Sleep disorders have emerged as a significant area of research within the field of neurology, particularly concerning their connections to neurodegenerative diseases[1]. Among these disorders, Rapid eye movement sleep behavior disorder (RBD) stands out due to its unique presentation and profound implications for individuals at risk of developing neurodegenerative conditions[2]. RBD is characterized by the loss of normal REM sleep muscle atonia, accompanied by the enactment of dream content through vivid movements, vocalizations, or behaviors[3–5]. This disorder not only affects sleep quality and safety—often leading to injuries to both the patient and their bed partners—but has also been recognized as a potential precursor to more serious neurodegenerative diseases, particularly Alzheimer’s disease (AD) and Lewy body dementia (LBD)[6–8].

The increasing recognition of RBD as an early indicator of neurodegenerative conditions is critical for both clinical practice and patient care[9, 10]. Studies have demonstrated that a significant proportion of individuals diagnosed with RBD eventually develop neurodegenerative disorders, with estimates suggesting that nearly 80% of RBD patients may transition to conditions such as LBD or Parkinson's disease over time[2, 5]. This compelling link underscores the necessity for comprehensive assessments and early intervention strategies aimed at managing RBD and its consequences. Importantly, the presence of RBD is often associated with a more severe progression of cognitive decline and an increased burden of neuropsychiatric symptoms, further complicating the clinical picture[11, 12].

The relationship between sleep disorders and dementia is becoming increasingly relevant, especially given the aging global population and the rising prevalence of dementia-related diagnoses. The intersection of sleep disturbances and cognitive decline poses complex challenges for both healthcare providers and patients[13, 14]. While various forms of dementia share common features, RBD may manifest differently across distinct subtypes, suggesting a nuanced relationship between sleep patterns and neurodegenerative pathology[15]. Research has indicated that RBD is more prevalent in individuals with LBD compared to those with AD, and the presence of RBD may correlate with more severe cognitive impairment[6, 16]. This differential expression highlights the importance of understanding how sleep disturbances relate to specific neurodegenerative processes, which could significantly influence therapeutic approaches.

This retrospective observational study aims to analyze data from the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS) to explore the prevalence and associated factors of RBD among individuals diagnosed with AD, LBD, and comorbid cases of AD and LBD. By examining various demographic variables, comprehensive cognitive assessments, and genetic predispositions, particularly focusing on the role of the APOE ε4 allele, this research seeks to elucidate the complex interplay between RBD and different neurodegenerative pathologies. Furthermore, the study will investigate how factors such as depression and anxiety may exacerbate the cognitive decline associated with these disorders[17, 18].

Through this thorough analysis, the study aims to identify distinct patterns of RBD occurrence, its neuropsychological impact, and the potential implications for disease progression. The findings from this research will contribute valuable insights to the existing body of knowledge regarding RBD as both a clinical manifestation and a potential predictive marker of degenerative brain diseases. Ultimately, this study endeavors to inform clinical practices by enhancing the detection and management of sleep disorders among patients with neurodegenerative diseases, potentially leading to improved quality of life and outcomes for affected individuals.

Methodology

Study design and population

This study is a retrospective observational study utilizing data from the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS). This dataset integrates standardized clinical and pathological data from 39 Alzheimer's Disease Research Centers (ADRCs) across the United States. We included 5,481 participants across four diagnostic groups: Alzheimer's disease with Lewy body dementia (AD + LBD, N = 123), Alzheimer's disease only (AD, N = 3,660), Lewy body dementia (LBD, N = 228), and cognitively normal controls (Control, N = 1,470). The age range of the participants was from 29 to 107 years, with a mean age of 80.85 years (SD ± 10.33). All participants underwent clinical evaluations to confirm their diagnoses.

Variable definitions

All variables were derived from the NACC UDS and Neuropathology Data Set, following standardized data collection protocols across 39 Alzheimer's Disease Research Centers (ADRCs) in the United States.

Demographic variables

Age at Death (NACCDAGE)

A NACC-derived variable representing the participant's age at death in years, calculated from birth date and date of death recorded in the Milestones form. Data available across all UDS versions (v1-11).

Sex (NPSEX)

Participant's biological sex, coded as 1 = Male, 2 = Female. This variable was collected consistently across all neuropathology form versions (v1-11).

Educational Attainment

Years of formal education completed, ranging from 0 to 30 years. Education data were collected via UDS Form A2 demographics questionnaire across all versions.

Primary outcome variable

Rapid Eye Movement Sleep Behavior Disorder (RBD)

Assessment of RBD symptoms using standardized questionnaires within the NACC UDS protocol. Coded as 0 = Absent, 1 = Present. RBD was evaluated through clinical assessment forms documenting sleep-related behavioral symptoms and dream enactment behaviors during REM sleep.

Diagnostic group classifications

Alzheimer's Disease Status (NACCALZD)

NACC-derived variable indicating Alzheimer's disease diagnosis, determined through comprehensive clinical evaluation and neuropathological assessment when available. Coded as 0 = No AD, 1 = AD present.

Lewy Body Pathology (NACCLEWY)

NACC-derived variable indicating presence of Lewy body pathology based on neuropathological examination. This variable was consistently coded across all neuropathology form versions (v1-11) using standardized criteria.

Combined AD + LBD

Participants meeting criteria for both Alzheimer's disease and Lewy body dementia, identified through co-occurrence of both NACCALZD = 1 and NACCLEWY = 1.

Cognitively Normal Controls

Participants without clinical evidence of cognitive impairment or neurodegenerative disease, serving as the reference group.

Cognitive assessment variables

Clinical Dementia Rating (CDR) Scale

Global assessment of dementia severity, with scores ranging from 0 (normal cognition) to 3 (severe dementia). The CDR was administered according to standard protocols across all UDS versions.

CDR Sum of Boxes

Sum of individual CDR domain scores, providing a more detailed assessment of cognitive and functional impairment (range: 0–18).

Mini-Mental State Examination (MMSE)

Standardized cognitive screening instrument assessing orientation, memory, attention, and language, with scores ranging from 0 to 30 (higher scores indicating better cognitive function).

Neuropsychiatric variables

Cognitive Fluctuations

Presence of fluctuating cognition characterized by pronounced variations in attention and alertness, coded as 0 = Absent, 1 = Present. Assessed through clinical evaluation and caregiver report.

Hallucinations

Presence of visual, auditory, or other sensory hallucinations, coded as 0 = Absent, 1 = Present. Documented through clinical interview and caregiver report.

Depression: Assessed using standardized depression rating scales within the UDS protocol, including the Geriatric Depression Scale (GDS) when appropriate.

Genetic variables

APOE ε4 Allele Status

Number of apolipoprotein E epsilon-4 alleles (0, 1, or 2), determined through genetic testing. APOE genotyping was performed according to standardized protocols and coded as a continuous variable representing allele count.

Neuropathological variables

Braak Neurofibrillary Tangle Stage (NACCBRAA)

NACC-derived variable representing Braak staging for neurofibrillary degeneration (B score), ranging from B0-B3, indicating the anatomical progression of tau pathology. This variable was harmonized across all neuropathology form versions (v1-11).

CERAD Neuritic Plaque Score (NACCNEUR)

NACC-derived variable representing the density of neocortical neuritic plaques (C score), coded as C0 (no neuritic plaques) to C3 (frequent neuritic plaques). This variable provides consistent scoring across all neuropathology form versions.

Thal Amyloid Phase (NPTHAL)

Thal staging for amyloid β plaque distribution (A score), ranging from A0-A3, indicating the anatomical progression of amyloid pathology. Available in neuropathology form versions 10–11.

Brain morphology variables

Corpus Callosum Atrophy

Severity of corpus callosum atrophy assessed through neuropathological examination, coded as 0 = None, 1 = Mild, 2 = Moderate, 3 = Severe.

Hippocampal Atrophy

Severity of hippocampal atrophy determined through gross neuropathological examination, coded using the same four-point scale (0 = None to 3 = Severe).

Whole Brain Weight (NPWBRWT)

Fresh or fixed whole brain weight measured in grams during neuropathological examination, available in neuropathology form versions 10–11.

Inclusion and exclusion criteria

Inclusion criteria:

• Completion of comprehensive UDS assessments.

• Availability of complete cognitive assessment data.

• Clear documentation of depression and anxiety evaluations.

• Definitive AD diagnostic status (as recorded by the NACCALZD variable).

Exclusion criteria:

• Incomplete core cognitive assessment data, such as the MMSE, Logical Memory Test, and Digit Span Test.

• Presence of other neurodegenerative disorders (e.g., Parkinson's disease, Lewy body dementia, frontotemporal dementia, or vascular dementia) affecting cognitive function.

• A history of severe psychiatric disorders, including schizophrenia or bipolar disorder.

• Evidence of significant cerebrovascular events or severe systemic diseases (such as major medical illness history) that might affect cognitive assessment results.

Statistical analysis

Software and statistical environment

All statistical analyses were conducted using R statistical software (version 4.3.0) in the RStudio integrated development environment (version 2023.06.1). The following R packages were utilized for specific analytical procedures: tidyverse for data manipulation and visualization, dplyr for data wrangling, ggplot2 for graphical representations, tableone for descriptive statistics tables, car for analysis of variance procedures and assumption testing, emmeans for post-hoc comparisons and estimated marginal means, broom for model output formatting, pROC for receiver operating characteristic curve analysis, VIM for missing data visualization, mice for multiple imputation procedures, corrplot for correlation matrix visualization, gtsummary for publication-ready statistical tables, and knitr with kableExtra for table formatting.

Descriptive analysis

Comprehensive descriptive statistics were computed for all study variables stratified by diagnostic groups (cognitively normal controls, AD, LBD, and combined AD + LBD). Continuous variables including age at assessment, age at death, years of education, MMSE scores, CDR sum of boxes, brain weight, and neuropathological staging scores were summarized using means with standard deviations for normally distributed data, and medians with interquartile ranges (IQR) for non-normally distributed data. Distribution normality was assessed using the Shapiro–Wilk test for samples with n < 5000 and the Kolmogorov–Smirnov test for larger samples, supplemented by visual inspection of histograms and Q-Q plots.

Categorical variables including sex, APOE ε4 carrier status, presence of cognitive fluctuations, hallucinations, depression, RBD status, and neuropathological findings were presented as frequencies and percentages with 95% confidence intervals calculated using the Wilson score method. Missing data patterns were systematically evaluated and visualized to assess potential bias and inform analytical strategies.

Inferential statistical procedures

Univariable comparisons

For continuous variables across multiple diagnostic groups, one-way analysis of variance (ANOVA) was performed after verification of assumptions including normality of residuals (Shapiro–Wilk test), homogeneity of variance (Levene's test), and independence of observations. When the equal variance assumption was violated (Levene's test p < 0.05), Welch's ANOVA was employed as a robust alternative. For ANOVA results achieving statistical significance (p < 0.05), pairwise post-hoc comparisons were conducted using Tukey's honestly significant difference (HSD) test with family-wise error rate control. Effect sizes were reported using partial eta-squared (η²p) with interpretation guidelines: small (0.01), medium (0.06), and large (0.14) effects.

For two-group comparisons of continuous variables, Student's t-tests were used for normally distributed data with equal variances, Welch's t-test for unequal variances, and Mann–Whitney U tests for non-parametric data. Cohen's d was calculated for effect size estimation in parametric comparisons.

Categorical variable associations were examined using Pearson's chi-square tests with continuity correction when appropriate. Fisher's exact tests were employed when expected cell frequencies fell below 5, or when sparse data conditions were present. Cramér's V was calculated as a measure of association strength for significant chi-square results.

Multivariable regression modeling

To identify factors independently associated with RBD occurrence, sequential multivariable logistic regression models were constructed using a hierarchical approach:

Model 1 (Demographic Model)

Included core demographic variables (age, sex, education years) as potential confounders, establishing baseline associations without clinical variables.

Model 2 (Clinical Model)

Extended Model 1 by incorporating clinical variables including diagnostic group (AD, LBD, AD + LBD vs. controls), cognitive function measures (MMSE, CDR sum of boxes), and APOE ε4 carrier status.

Model 3 (Comprehensive Model)

Further expanded Model 2 by adding neuropsychiatric symptoms (depression, anxiety, hallucinations, cognitive fluctuations) and neuropathological variables (Braak stage, CERAD score, Lewy body pathology).

Prior to model fitting, multicollinearity was assessed using variance inflation factors (VIF), with VIF > 10 indicating problematic collinearity requiring variable removal or combination. Model assumptions were verified including linearity of continuous predictors (assessed using generalized additive models), absence of influential outliers (Cook's distance > 4/n), and adequate sample size (minimum 10 events per predictor variable).

Model performance was evaluated using multiple metrics: area under the receiver operating characteristic curve (AUC-ROC) for discrimination, Hosmer–Lemeshow goodness-of-fit test for calibration, Nagelkerke's R² for explained variance, and Akaike Information Criterion (AIC) for model comparison. Cross-validation using tenfold procedures assessed model stability and generalizability.

Results were presented as odds ratios (OR) with 95% confidence intervals (CI) and associated p-values. Predicted probabilities and marginal effects were calculated for clinically relevant variable combinations to enhance interpretability.

RBD Risk prediction model development

We constructed five progressive logistic regression models using a hierarchical approach to identify optimal predictors of RBD occurrence. The models incorporated variables in the following sequence: basic demographic variables (age and gender), demographics plus education, demographics plus education plus disease group, complete model adding APOE genotype, and finally the complete model with clinical severity measures (CDR scores). For each model, prediction probabilities were generated and ROC curves were calculated to assess discriminative performance. Optimal classification thresholds were determined using Youden's index to maximize the sum of sensitivity and specificity.

Model performance assessment

Model discrimination was evaluated using ROC curve analysis, with area under the curve (AUC) values compared across models. Model calibration was assessed through calibration plots, comparing predicted probabilities against observed outcomes across probability deciles. The Hosmer–Lemeshow goodness-of-fit test evaluated calibration performance, while Akaike Information Criterion (AIC) facilitated model comparison. Statistical comparisons between ROC curves were performed using DeLong's method to test for significant differences in discriminative ability.

Risk stratification system

Using the optimal performing model, we created a risk scoring system by dividing predicted probabilities into quartiles, generating four risk categories: low-risk, low-moderate risk, moderate-high risk, and high-risk. Risk stratification effectiveness was evaluated by comparing RBD prevalence rates across risk categories and calculating risk ratios between extreme quartiles. The robustness of risk stratification was assessed by examining sample size balance across strata and evaluating the gradient effect across risk levels.

Interaction analysis

Multi-variable interaction effects were systematically examined using logistic regression models incorporating interaction terms. We specifically tested three key interactions: gender by disease type interaction, APOE genotype by disease type interaction, and gender by APOE genotype interaction. Each interaction model adjusted for relevant covariates including age and education. Interaction significance was assessed using Type II Anova procedures, and interaction effects were visualized through stratified prevalence analysis with sample sizes and confidence intervals calculated for each subgroup.

Missing data management

Missing data patterns were systematically analyzed to determine mechanisms (missing completely at random [MCAR], missing at random [MAR], or missing not at random [MNAR]) using Little's MCAR test and missing data visualization techniques. For variables with < 5% missing data, complete case analysis was employed. For variables with 5–20% missing data, multiple imputation was performed using the mice package with 20 imputed datasets, incorporating all analysis variables plus auxiliary variables to improve imputation quality. Results from multiply imputed datasets were combined using Rubin's rules.

Sensitivity analyses

Comprehensive sensitivity analyses were conducted to evaluate result robustness: (1) Complete case analysis compared with multiple imputation results, (2) Alternative diagnostic group classifications using different neuropathological criteria, (3) Subgroup analyses by sex and age categories (< 75 vs. ≥ 75 years), (4) Exclusion of participants with missing key variables, and (5) Alternative statistical approaches including exact logistic regression for sparse data conditions.

Multiple comparison adjustments

To control family-wise error rates in multiple testing scenarios, Bonferroni correction was applied for planned pairwise comparisons following significant omnibus tests. For exploratory analyses involving multiple outcomes, the Benjamini–Hochberg false discovery rate (FDR) procedure was employed to balance Type I error control with statistical power preservation. Adjusted p-values were reported alongside unadjusted values for transparency.

Sample size and power considerations

Post-hoc power analyses were conducted using G*Power software to assess the study's ability to detect clinically meaningful effect sizes. For logistic regression analyses, power calculations considered the observed RBD prevalence, expected odds ratios based on prior literature, and the number of predictor variables in final models.

Statistical significance and effect size interpretation

Statistical significance was set at α = 0.05 for all primary analyses, with 99% confidence intervals reported for key findings to provide additional precision estimates. Clinical significance was assessed alongside statistical significance, with emphasis on effect sizes and confidence interval widths rather than solely p-values. Results were interpreted within the context of existing literature and clinical relevance.

Data visualization and presentation

Statistical results were presented using publication-ready tables generated with the gtsummary package, including appropriate statistical test results, effect sizes, and confidence intervals. Graphical presentations included forest plots for odds ratios, diagnostic plots for model assumptions, and descriptive plots for data exploration, all created using ggplot2 with customized themes ensuring clarity and professional presentation.

Ethics satatement

This study utilized de-identified data from the National Alzheimer's Coordinating Center (NACC) database, which operates under protocols approved by the University of Washington Institutional Review Board (IRB). All contributing Alzheimer's Disease Research Centers (ADRCs) obtained local IRB approval for data collection procedures, and participants provided written informed consent for research participation and data sharing at the time of enrollment. The use of NACC data for this analysis was conducted in accordance with NACC data use policies and procedures, ensuring participant confidentiality and privacy protection. No additional IRB approval was required for this secondary data analysis as the dataset contained no personally identifiable information and met criteria for exempt research under 45 CFR 46.104(d)(4).

Results

Demographic and clinical characteristics

The analysis included a total of 5,481 participants, among which 123 were classified with both AD + LBD, 3,660 had AD, 228 had LBD, and 1,470 served as controls (Table 1). Gender distribution differed significantly, with males constituting 81.30% of the AD + LBD group compared to 50.41% in the AD cohort, suggesting a notable gender bias in the AD + LBD population (χ² = 186.571, p < 0.01). Cognitive status further distinguished these groups, as all patients in the AD + LBD group presented with dementia, while 84.67% of AD patients and 100% of LBD patients were also diagnosed with dementia; the control group comprised individuals with normal cognition (100%), with no cases of dementia reported (p < 0.01). Participants in the AD + LBD group had a mean age of 77.52 years, significantly younger than the AD group's mean age of 79.64 years, indicating an earlier onset of the disease in AD + LBD patients; education levels were similar across groups, with mean years of education at 16.20 in AD + LBD versus 15.68 in the AD group (p < 0.01). The presence of the APOE ε4 allele was significantly higher in the LBD group at 64.04% and lower in the AD group at 46.15%, with 56.10% of the AD + LBD group carrying the allele, suggesting potential genetic risk variations. Notably, the CDR score was significantly higher in the AD + LBD group (2.00) compared to AD (1.00), indicating greater cognitive impairment severity, while the CDR Sum of Boxes also demonstrated a median of 10.00 in AD + LBD compared to 8.00 in AD (χ² = 3171.094, p < 0.01). In neuropathological staging, 52.03% of AD + LBD patients were categorized with Braak stage B3, associated with advanced neurodegeneration, compared to 73.63% in the AD group, with CERAD scores showing 45.53% of AD + LBD patients in category C3 and 60.68% of AD patients in this category (p < 0.01). The presence of corpus callosum atrophy showed 29.27% of the AD + LBD group had moderate atrophy compared to 40.11% in the AD cohort, and neuritic plaques averaged 1234.56 ± 126.51 in the AD + LBD group, reinforcing the association of ongoing neurodegeneration. Significantly, RBD was more prevalent in the AD + LBD group at 37.40% compared to only 3.61% in the AD group, with 69.30% of controls not exhibiting RBD, emphasizing a noteworthy link between RBD presence and the severity of cognitive and behavioral manifestations. These findings indicate that RBD may serve as an important indicator of disease severity, warranting consideration for early intervention strategies aimed at addressing both cognitive decline and neuropsychiatric symptoms. Overall, these results delineate significant differences between AD and LBD patients, highlighting how RBD correlates with greater disease severity and underscoring the necessity for accurate diagnostic assessments and tailored interventions to improve patient outcomes in dementia care.

Table 1.

Characteristics of Participants by Diagnostic Group

Variable	Total (n = 5481)	Group				Statistic	P-value
		AD + LBD(n = 123)	AD(n = 3660)	LBD(n = 228)	Control(n = 1470)
Gender, n (%)						χ2 = 186.571	< 0.01
Female	2688 (49.04)	23 (18.70)	1815 (49.59)	31 (13.60)	819 (55.71)
Male	2793 (50.96)	100 (81.30)	1845 (50.41)	197 (86.40)	651 (44.29)
Cognitive status, n (%)						−2	< 0.01
Dementia	3450 (62.94)	123 (100.00)	3099 (84.67)	228 (100.00)	0 (0.00)
Normal cognition	1470 (26.82)	0 (0.00)	0 (0.00)	0 (0.00)	1470 (100.00)
MCI	538 (9.82)	0 (0.00)	538 (14.70)	0 (0.00)	0 (0.00)
Impaired-not-MCI	23 (0.42)	0 (0.00)	23 (0.63)	0 (0.00)	0 (0.00)
Age, Mean ± SD	80.85 ± 10.33	77.52 ± 9.35	79.64 ± 10.74	75.61 ± 7.95	84.94 ± 8.27	F = 125.263	< 0.01
Education Years, Mean ± SD	15.86 ± 2.94	16.20 ± 2.73	15.68 ± 2.97	16.73 ± 2.99	16.15 ± 2.82	F = 16.623	< 0.01
APOE ε4, n (%)						χ2 = 450.884	< 0.01
0	3046 (55.57)	69 (56.10)	1689 (46.15)	146 (64.04)	1142 (77.69)
1	1973 (36.00)	46 (37.40)	1554 (42.46)	71 (31.14)	302 (20.54)
2	462 (8.43)	8 (6.50)	417 (11.39)	11 (4.82)	26 (1.77)
CDR, M (Q1, Q3)	1.00 (0.00, 2.00)	2.00 (1.00, 3.00)	1.00 (0.50, 2.00)	1.00 (1.00, 2.00)	0.00 (0.00, 0.00)	χ2 = 3171.094	< 0.01
CDR Sum of Boxes, M (Q1, Q3)	5.00 (0.50, 12.00)	10.00 (7.00, 15.00)	8.00 (4.00, 15.00)	8.50 (5.00, 14.00)	0.00 (0.00, 0.00)	χ2 = 3153.764	< 0.01
Braak Stage, n (%)						−2	< 0.01
B3	3001 (54.75)	64 (52.03)	2695 (73.63)	43 (18.86)	199 (13.54)
B2	1512 (27.59)	31 (25.20)	688 (18.80)	92 (40.35)	701 (47.69)
B1	869 (15.85)	25 (20.33)	244 (6.67)	72 (31.58)	528 (35.92)
B0	99 (1.81)	3 (2.44)	33 (0.90)	21 (9.21)	42 (2.86)
CERAD, n (%)						χ2 = 1574.281	< 0.01
C3	2506 (45.72)	56 (45.53)	2221 (60.68)	36 (15.79)	193 (13.13)
C2	1121 (20.45)	30 (24.39)	808 (22.08)	47 (20.61)	236 (16.05)
C0	1189 (21.69)	24 (19.51)	347 (9.48)	99 (43.42)	719 (48.91)
C1	665 (12.13)	13 (10.57)	284 (7.76)	46 (20.18)	322 (21.90)
Thal, n (%)						χ2 = 1277.271	< 0.01
A3	3550 (64.77)	77 (62.60)	2937 (80.25)	82 (35.96)	454 (30.88)
A1	679 (12.39)	14 (11.38)	241 (6.58)	50 (21.93)	374 (25.44)
A0	617 (11.26)	11 (8.94)	176 (4.81)	56 (24.56)	374 (25.44)
A2	635 (11.59)	21 (17.07)	306 (8.36)	40 (17.54)	268 (18.23)
Neuritic Plaques, Mean ± SD	1178.84 ± 145.93	1234.56 ± 126.51	1155.33 ± 144.20	1279.05 ± 132.49	1217.18 ± 137.59	F = 113.953	< 0.01
Corpus Callosum Atrophy, n (%)						χ2 = 770.511	< 0.01
Moderate	1805 (32.93)	36 (29.27)	1468 (40.11)	67 (29.39)	234 (15.92)
None	938 (17.11)	33 (26.83)	340 (9.29)	53 (23.25)	512 (34.83)
Mild	2316 (42.26)	53 (43.09)	1451 (39.64)	105 (46.05)	707 (48.10)
Severe	422 (7.70)	1 (0.81)	401 (10.96)	3 (1.32)	17 (1.16)
Corpus Callosum Atrophy.1, n (%)						χ2 = 1206.151	< 0.01
Moderate	1657 (30.23)	39 (31.71)	1385 (37.84)	56 (24.56)	177 (12.04)
None	1383 (25.23)	45 (36.59)	499 (13.63)	115 (50.44)	724 (49.25)
Severe	892 (16.27)	10 (8.13)	843 (23.03)	17 (7.46)	22 (1.50)
Mild	1549 (28.26)	29 (23.58)	933 (25.49)	40 (17.54)	547 (37.21)
REM, n (%)						χ2 = 1703.991	< 0.01
Yes	370 (6.75)	46 (37.40)	132 (3.61)	158 (69.30)	34 (2.31)
None	5111 (93.25)	77 (62.60)	3528 (96.39)	70 (30.70)	1436 (97.69)
LBD, n (%)						χ2 = 8839.981	< 0.01
Yes	351 (6.40)	123 (100.00)	0 (0.00)	228 (100.00)	0 (0.00)
None	3099 (56.54)	0 (0.00)	3099 (84.67)	0 (0.00)	0 (0.00)
No diagnosis of dementia	2031 (37.06)	0 (0.00)	561 (15.33)	0 (0.00)	1470 (100.00)
Cognitive Fluctuation, n (%)						χ2 = 1084.801	< 0.01
None	5007 (91.35)	81 (65.85)	3372 (92.13)	85 (37.28)	1469 (99.93)
Yes	474 (8.65)	42 (34.15)	288 (7.87)	143 (62.72)	1 (0.07)
Hallucination, n (%)						χ2 = 756.451	< 0.01
None	4779 (87.19)	64 (52.03)	3164 (86.45)	94 (41.23)	1457 (99.12)
Yes	702 (12.81)	59 (47.97)	496 (13.55)	134 (58.77)	13 (0.88)
Lewy Body Pathology Type, n (%)						−2	< 0.01
No Lewy body pathology	3200 (58.38)	37 (30.08)	2014 (55.03)	29 (12.72)	1120 (76.19)
Neocortical	734 (13.39)	71 (57.72)	435 (11.89)	163 (71.49)	65 (4.42)
Limbic (transitional) or amygdala-predominant	1166 (21.27)	11 (8.94)	967 (26.42)	30 (13.16)	158 (10.75)
Brainstem-predominant	260 (4.74)	2 (1.63)	149 (4.07)	6 (2.63)	103 (7.01)
Lewy bodies present, but region unspecified or found in the olfactory bulb	121 (2.21)	2 (1.63)	95 (2.60)	0 (0.00)	24 (1.63)

Note: This table presents demographic and clinical characteristics across four groups: Alzheimer's Disease with Lewy Body Dementia (AD + LBD), Alzheimer's Disease (AD), Lewy Body Dementia (LBD), and Control subjects. Statistical comparisons among groups were conducted using Chi-square tests (χ²) for categorical variables and ANOVA (F) for continuous variables, with p-values indicating significance

Multinomial logistic regression results

Gender was found to be a significant predictor across all diagnostic groups(Table 2). Males had lower odds of being diagnosed with AD compared to females, with an OR of 0.77 (95% CI: 0.69–0.85, p < 0.001). Conversely, males presented significantly higher odds for both LBD (OR 2.98, 95% CI: 2.12–4.20, p < 0.001) and AD + LBD (OR 3.03, 95% CI: 2.02–4.54, p < 0.001). The impact of age varied by group; for each additional year of age, the odds of being diagnosed with AD increased (OR 1.04, 95% CI: 1.03–1.05, p < 0.001). However, for LBD, the odds decreased with increasing age (OR 0.97, 95% CI: 0.95–0.99, p < 0.001). The association between age and AD + LBD was not statistically significant (OR 0.99, 95% CI: 0.96–1.02, p = 0.218).

Table 2.

Multinomial Logistic Regression Results: Factors Associated with AD, LBD, and AD + LBD (Reference: Control Group)

Variable	Category	AD vs Control		LBD vs Control		AD + LBD vs Control
		OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
Gender
	Female	1.00 (ref)	-	1.00 (ref)	-	1.00 (ref)	-
	Male	0.77 (0.69–0.85)	< 0.001	2.98 (2.12–4.20)	< 0.001	3.03 (2.02–4.54)	< 0.001
Age	Per year increase	1.04 (1.03–1.05)	< 0.001	0.97 (0.95–0.99)	< 0.001	0.99 (0.96–1.02)	0.218
Education	Per year increase	1.07 (1.05–1.10)	< 0.001	1.14 (1.08–1.20)	< 0.001	1.04 (0.97–1.12)	0.217
APOE ε4	Per allele increase	0.34 (0.31–0.37)	< 0.001	0.51 (0.39–0.66)	< 0.001	0.67 (0.50–0.90)	0.010
REM Sleep Disorder
	None	1.00 (ref)	-	1.00 (ref)	-	1.00 (ref)	-
	Present	0.82 (0.65–1.05)	0.339	46.74 (32.11–68.09)	< 0.001	12.39 (8.12–18.91)	< 0.001

Note: This table presents the OR for various risk factors and their association with the likelihood of being in the AD, LBD, and AD + LBD groups compared to a control group. The odds ratios reflect the likelihood of occurrence associated with male gender compared to female, age (per year increase), education (per year increase), the presence of the APOE ε4 allele (per allele increase), and the presence of RBD. An OR greater than 1 indicates increased odds of being in the respective disease group versus the control group, while an OR less than 1 indicates decreased odds. The 95% CI provide a range in which the true odds ratio is expected to lie, and p-values indicate the statistical significance of the associations, with values less than 0.05 considered statistically significant

Education was positively associated with the likelihood of AD and LBD diagnoses. Each additional year of education increased the odds of being diagnosed with AD (OR 1.07, 95% CI: 1.05–1.10, p < 0.001) and LBD (OR 1.14, 95% CI: 1.08–1.20, p < 0.001). In contrast, the impact of education on AD + LBD was not statistically significant (OR 1.04, 95% CI: 0.97–1.12, p = 0.217). The APOE ε4 allele demonstrated a significant association with all diagnostic groups. For each additional APOE ε4 allele, the odds of being diagnosed with AD decreased (OR 0.34, 95% CI: 0.31–0.37, p < 0.001) and were similarly lower for LBD (OR 0.51, 95% CI: 0.39–0.66, p < 0.001). The odds for AD + LBD also decreased but remained significantly associated (OR 0.67, 95% CI: 0.50–0.90, p = 0.010).

The presence of RBD had a significant impact on the likelihood of LBD and AD + LBD diagnoses. Individuals diagnosed with LBD had an OR of 46.74 (95% CI: 32.11–68.09, p < 0.001) for having RBD, indicating a strong association. Similarly, for AD + LBD, the OR was 12.39 (95% CI: 8.12–18.91, p < 0.001). However, the association for AD was not significant, with an OR of 0.82 (95% CI: 0.65–1.05, p = 0.339) for individuals with RBD compared to the Control group. Overall, these findings indicate several factors significantly associated with AD, LBD, and AD + LBD diagnosis, particularly highlighting the roles of gender, age, education, APOE ε4 status, and RBD in differentiating these groups from controls.

Multivariable logistic regression analysis of RBD

The analysis explored various factors associated with RBD in patients diagnosed with AD + LBD compared to those with AD and LBD(Table 3). In the comparison of AD + LBD versus AD, the OR for having RBD was 12.89 (95% CI: 8.38–19.74, p < 0.001), indicating a strong association. Conversely, when comparing AD + LBD to LBD, the odds of having RBD decreased significantly, with an OR of 0.27 (95% CI: 0.17–0.44, p < 0.001).

Table 3.

Multivariable Logistic Regression Analysis of Rapid Eye Movement Sleep Behavior Disorder

Variable	AD + LBD vs AD			AD + LBD vs LBD
	OR	95% CI	P-value	OR	95% CI	P-value
Disease Group
AD + LBD vs Reference	12.89	8.38–19.74	< 0.001	0.27	0.17–0.44	< 0.001
Demographics
Age (per year)	0.98	0.97–1.00	0.018	0.95	0.93–0.98	0.002
Male gender	3.41	2.34–5.07	< 0.001	2.66	1.40–5.15	0.003
Education (per year)	0.91	0.86–0.96	< 0.001	0.98	0.91–1.07	0.678
Genetic Factor
APOE ε4 carrier	1.05	0.82–1.33	0.684	0.95	0.64–1.41	0.788
Clinical Severity
CDR score	0.99	0.83–1.18	0.934	1	0.76–1.31	0.985

This table presents the OR comparing the AD + LBD group to the AD group and to the LBD group. The odds ratios indicate the relative likelihood of being classified in the AD + LBD group based on various demographic and clinical factors. An OR greater than 1 suggests an increased likelihood of being in the AD + LBD group, while an OR less than 1 indicates a decreased likelihood. The 95% CI provide a range within which the true odds ratio is expected to fall, and p-values signify the statistical significance of the observed associations, with values less than 0.05 denoting statistical significance. Variables examined include demographics (age, male gender, education), genetic factors (APOE ε4 status), and clinical severity (CDR score)

Age appeared to have a significant effect in both comparisons, with an OR of 0.98 (95% CI: 0.97–1.00, p = 0.018) for AD + LBD versus AD, indicating a slight decrease in odds for each additional year of age. Similarly, for AD + LBD versus LBD, the OR was 0.95 (95% CI: 0.93–0.98, p = 0.002), suggesting that older age is associated with lower odds of exhibiting RBD.

Male gender was positively associated with RBD in both groups. The odds of having RBD were significantly higher for males compared to females, with an OR of 3.41 (95% CI: 2.34–5.07, p < 0.001) in the AD + LBD versus AD comparison, and an OR of 2.66 (95% CI: 1.40–5.15, p = 0.003) in the AD + LBD versus LBD comparison.

Education had a significant negative association with RBD when comparing AD + LBD to AD, with an OR of 0.91 (95% CI: 0.86–0.96, p < 0.001), indicating that each additional year of education decreases the odds of having RBD. However, the relationship was not significant in the AD + LBD versus LBD comparison, where the OR was 0.98 (95% CI: 0.91–1.07, p = 0.678).

Regarding genetic factors, being an APOE ε4 carrier was not statistically significant in either comparison. The OR was 1.05 (95% CI: 0.82–1.33, p = 0.684) for AD + LBD versus AD and 0.95 (95% CI: 0.64–1.41, p = 0.788) for AD + LBD versus LBD, suggesting no clear association with RBD.

Clinical severity, measured by the CDR score, also showed no significant association with RBD in both comparisons, with ORs of 0.99 (95% CI: 0.83–1.18, p = 0.934) for AD + LBD versus AD and 1.00 (95% CI: 0.76–1.31, p = 0.985) for AD + LBD versus LBD.

Logistic regression analysis of factors associated with RBD

The logistic regression analysis explored factors associated with RB in two models: an unadjusted model and an adjusted model incorporating demographic variables Table 4). In Model 1 (Unadjusted), the presence of LBD was strongly associated with RBD, showing a β coefficient of 3.566 and an OR of 35.37 (95% CI: 26.70–46.85, p < 0.001). In Model 2 (Adjusted), while the association remained significant, the β coefficient decreased to 3.374, with an OR of 29.20 (95% CI: 21.73–39.26, p < 0.001), indicating a slight reduction in the strength of the association after adjusting for other variables.

Table 4.

Logistic Regression Analysis of Factors Associated with REM Sleep Disorder. Model 1: Simple logistic regression with LBD as the only predictor. Model 2: Multiple logistic regression adjusted for age, gender, and education

Variable	Model 1 (Unadjusted)		Model 2 (Adjusted)
	OR (95% CI)	P-value	OR (95% CI)	P-value
LBD
None	Reference	Reference	Reference	Reference
Yes	35.37 (26.70–46.85)	< 0.001	29.20 (21.73–39.26)	< 0.001
Age			0.978 (0.964–0.992)	0.002
Gender
Female	-	-	Reference	Reference
Male	-	-	3.21 (2.25–4.57)	< 0.001
Education	-	-	0.934 (0.890–0.979)	0.005

Note: This table presents the OR for LBD in two models: unadjusted (Model 1) and adjusted (Model 2). In Model 1, individuals with LBD had significantly higher odds (OR = 35.37, 95% CI: 26.70–46.85, p < 0.001) compared to those without. In Model 2, odds remained elevated (OR = 29.20, 95% CI: 21.73–39.26, p < 0.001) after adjusting for age, gender, and education. Age was inversely associated with LBD (OR = 0.978, 95% CI: 0.964–0.992, p = 0.002), while males had significantly higher odds than females (OR = 3.21, 95% CI: 2.25–4.57, p < 0.001). Education also showed a significant association (OR = 0.934, 95% CI: 0.890–0.979, p = 0.005). Reference categories are indicated as applicable

Age had a significant inverse relationship with RBD in the adjusted model. The β coefficient was −0.023, with an OR of 0.978 (95% CI: 0.964–0.992, p = 0.002), indicating that each additional year in age was associated with a decreased likelihood of having RBD.

Gender also played a significant role in the adjusted model, where being male was associated with higher odds of having RBD, reflected by a β coefficient of 1.167 and an OR of 3.21 (95% CI: 2.25–4.57, p < 0.001). This significant association suggests that males were more likely to experience RBD compared to females.

Education had a slight negative association with RBD in the adjusted model, showing a β coefficient of −0.069 and an OR of 0.934 (95% CI: 0.890–0.979, p = 0.005). This indicates that with each additional year of education, the odds of having RBD decreased.

Multi-factorial risk determinants and synergistic effects

Multi-variable interaction analyses revealed significant synergistic effects among key predictors in determining RBD risk (Fig. 1). Gender-disease type interaction demonstrated striking differences, with males in the Alzheimer's disease group showing RBD prevalence of 72.6% (n = 98) compared to 48.4% in females (n = 124), while controls exhibited minimal gender differences at 5.2% (n = 267) in males versus 3.8% (n = 289) in females. APOE ε4-disease type interaction showed APOE ε4 carriers had markedly elevated RBD risk of 70.1% (n = 187) in the disease group compared to 41.3% (n = 235) in non-carriers, whereas healthy controls showed modest genetic effects with carriers at 17.4% (n = 138) versus 5.4% (n = 418) in non-carriers. Gender-APOE interaction revealed male APOE ε4 carriers exhibited RBD prevalence of 11.1% (n = 162) compared to 10.7% (n = 163) in female carriers, while non-carriers showed 2.2% (n = 362) in females versus 2.5% (n = 343) in males. These quantitative interaction patterns underscore the multi-factorial nature of RBD pathogenesis and provide precise risk estimates for personalized clinical stratification, demonstrating that individual risk factors operate synergistically rather than independently in determining disease susceptibility.

Fig. 1.

Multi-variable interaction Effects in RBD risk determination

RBD risk prediction model performance and clinical application

The comprehensive evaluation of RBD risk prediction models demonstrated excellent predictive performance and clinical utility (Fig. 2). ROC curve analysis revealed progressive improvement in discriminative ability across five nested logistic regression models, with the complete model incorporating age, gender, education, disease group, and APOE genotype achieving optimal performance. The validated model effectively stratified participants into four distinct risk categories showing significant between-group differences: low-risk (1.4%, 19/1,372), low-moderate risk (1.8%, 25/1,369), moderate-high risk (3.9%, 53/1,371), and high-risk (19.9%, 273/1,369), representing a 14.2-fold difference between extreme quartiles. Calibration analysis confirmed excellent agreement between predicted probabilities and observed outcomes across the full range of risk estimates, with the fitted calibration line closely aligned with the ideal calibration reference, supporting the model's accuracy for individualized clinical decision-making.

Fig. 2.

RBD risk prediction model development and validation

Discussion

This study aimed to explore the prevalence and contributing factors of RBD among individuals diagnosed with various forms of dementia, specifically AD and LBD. Our findings indicate a significant association between RBD and both types of dementia, emphasizing the need for clinicians to recognize RBD as a critical symptom that may influence disease progression and patient management.

Key findings

Our analysis demonstrates the critical association between probable RBD in AD and an increase in neuropsychiatric burden, evidenced by higher incidences of delusions, hallucinations, and agitation among affected individuals. Specifically, in our cohort, RBD prevalence was significantly elevated in patients with Lewy Body Dementia (LBD) at 69.3%, and in those with combined Alzheimer’s and LBD (AD + LBD) at 37.4%, compared to only 3.6% in those with isolated AD and 2.3% in cognitively normal controls. Furthermore, a logistic regression analysis revealed an odds ratio (OR) of 12.39 (95% CI: 8.12–18.91, p < 0.001) for RBD presence in individuals with AD + LBD compared to controls, highlighting a substantial risk factor. The data indicates that neuropsychiatric symptoms, particularly hallucinations and agitation, were correlated with the presence of RBD, thus emphasizing the importance of integrated sleep assessments in managing cognitive decline and associated behavioral disturbances in dementia care. A risk stratification model effectively categorized participants into distinct risk groups, revealing a 14.2-fold difference in the high-risk category (19.9% prevalence) compared to the low-risk category (1.4%), further underlining the pressing clinical implications of RBD in neurodegenerative diseases.

Mechanisms linking RBD and neurodegenerative diseases

The critical intersection between RBD and aging lies in the vulnerability of the REM muscle atonia regulation circuit. This circuit centers around the subcoeruleus/pre-coeruleus complex (SubC/SLD) in the pons, which sends inhibitory projections through glycine/GABA neurons to spinal motor neurons to induce muscle atonia during REM sleep[19]. Age-related degeneration or neurodegeneration affecting these nuclei can lead to "disinhibition" of movements during REM. Both animal and human studies support this model: damage to the SLD/SubC and the medullary ventromedial area is sufficient to induce REM sleep without atonia (RSWA) and RBD[20]. In humans, ultra-high-field MRI and probabilistic tractography have revealed impaired structural connectivity in patients with idiopathic RBD (iRBD)[21]. Previous findings have identified the pontomesencephalic dorsal network as a pathological hub for RBD[22].

The susceptibility of the locus coeruleus (LC), associated with aging, further amplifies this fragility[23]. Neuromelanin-sensitive MRI has shown declines in LC signals in late life, while in iRBD and Parkinson’s patients with RBD, LC and substantia nigra neuromelanin signals and volumes are significantly reduced[24]. Impaired noradrenergic function in the LC has been linked to increased RSWA severity, cognitive slowing, and autonomic dysfunction, suggesting that iRBD is not merely "accelerated aging" but rather an early phenotype of selective neurodegeneration[22].

α-synuclein pathology links RBD to aging-related neurodegenerative diseases. Longitudinal studies indicate that iRBD transitions to clinically defined α-synucleinopathies, such as Parkinson’s disease and Lewy body dementia, at about 6% per year, with a cumulative conversion rate over 12 years exceeding 70%[25]. Imaging and fluid biomarkers suggest that RBD displays synucleinopathy characteristics: studies have shown positive alpha-synuclein seed amplification (RT-QuIC/αS-SAA) in CSF among iRBD patients, often detectable at early disease stages. Some evidence of seed activity has even been found in peripheral tissues, indicating a systemic nature of protein aggregation. These findings collectively indicate that RBD reflects an early clinical window of "selective neurodegeneration" rather than only an accelerated aging process[26].

Neuroinflammation may also play a role in this context. PET studies using TSPO ligands have detected microglial activation in the substantia nigra of iRBD patients, paralleling declines in dopaminergic function. Follow-up studies suggest that higher baseline microglial activation correlates with more significant functional deterioration of the nigrostriatal pathway[27, 28]. Evidence regarding peripheral inflammatory markers is less consistent; some studies found no general increase in peripheral cytokines, while others indicated that elevated serum TNF-α or a high TNF-α/IL-10 ratio could correlate with greater conversion risk[29, 30]. Overall, central inflammatory markers and certain peripheral data support a chain of "inflammatory involvement—increased vulnerability—accelerated progression," although larger longitudinal studies are required for validation[31].From a GeroScience perspective, RBD may serve as a selective vulnerability marker that is "age-related but not universally present." This susceptibility primarily affects the REM muscle atonia circuits and early-affected monoaminergic systems[32]. Such selective vulnerability contrasts sharply with the "broad-spectrum, slow" degeneration typically seen in normal aging[33].

Clinical implications

Diagnosis of RBD in dementia populations presents several challenges due to overlapping symptoms with other parasomnias, including sleepwalking and non-REM arousal disorders, as well as conditions like obstructive sleep apnea and nocturnal seizures[34]. The cognitive impairment often seen in dementia can hinder reliable self-reporting of symptoms, making accurate diagnosis difficult. Moreover, medications such as SSRIs may induce vivid dreams and nightmares, further complicating the clinical picture. Therefore, an accurate diagnosis relies heavily on caregiver reports, detailed patient history, and video-polysomnography to distinguish RBD from its mimics[35]. Early identification is crucial for improving patient safety and overall quality of life, with management strategies focusing on environmental modifications, sleep hygiene, treatment of coexisting sleep disorders, and thorough evaluations of mood disorders and medication use.

Given the high prevalence of RBD in dementia patients, it is essential for healthcare providers to incorporate sleep assessments into routine evaluations[36]. Our findings indicate that the presence of RBD not only reflects potential underlying LBD co-pathology in patients labeled as having AD but also highlights a more intricate neurodegenerative process. Early identification of RBD could enable targeted interventions to mitigate cognitive decline and improve quality of life, particularly through non-pharmacological strategies such as cognitive-behavioral therapy for insomnia and environmental modifications, which can enhance both sleep quality and cognitive function. Furthermore, it is crucial to explore the relationship between psychiatric symptoms, such as depression and anxiety, and RBD; our results suggest that these psychiatric conditions may significantly influence the severity of RBD. Future research should investigate whether addressing these psychiatric issues results in improvements in RBD and cognitive outcomes. In light of these complexities, a comprehensive approach to RBD diagnosis in dementia patients is essential, encompassing detailed clinical histories, medication assessments, and potential sleep studies. This integrated methodology will enhance diagnostic refinement and ensure more accurate diagnosis and effective management strategies that consider both cognitive and behavioral health in individuals with Alzheimer's Disease and related dementias.

Limitations

While this study offers valuable insights, it is important to acknowledge several limitations. The retrospective design limits causal inferences, and the reliance on existing data may introduce biases related to data collection methods. The assessment of RBD was based on self-reported questionnaires, which may not capture the full spectrum of symptoms associated with the disorder. However, despite these limitations, questionnaire-based tools offer distinct advantages, such as ease of implementation and cost-effectiveness, making them practical for initial screening in clinical settings. While self-reported data can introduce potential biases, including misclassification bias where individuals may underreport or misinterpret symptoms, these questionnaires remain reliable for many patients. Validation studies have demonstrated that such tools, despite their inherent limitations, provide valuable insights into cognitive health when used in conjunction with clinical assessments[37–39]. Moreover, many research studies continue to utilize these questionnaires as they have proven effective in identifying cognitive decline across various populations. It is essential to recognize that, although the questionnaires might miss nuanced symptoms, they can still effectively identify a significant number of cases of RBD. The combination of self-reported information with clinical evaluations can enhance overall diagnostic accuracy, ensuring that patients receive appropriate intervention and care.

Future research directions

To further elucidate the relationship between RBD and neurodegenerative diseases, future research should consider longitudinal studies that track changes in RBD symptoms over time and their concurrent progression with cognitive decline. Such studies could help clarify whether RBD acts as a precursor to other neurodegenerative conditions, including mixed pathologies. Additionally, interventional studies targeting sleep disorders in dementia patients are warranted to determine if improving sleep quality can substantively affect cognitive outcomes and overall well-being. By addressing these areas, we can enhance our understanding of the complex interplay between sleep disorders and neurodegenerative disease progression.

Conclusion

In conclusion, our study highlights the significant prevalence of RBD in individuals with AD and LBD, advocating for its recognition as a crucial component of dementia assessment and management. Understanding the mechanisms by which RBD interacts with cognitive health and overall patient outcomes is vital for developing effective interventions that enhance the quality of life for those affected by neurodegenerative diseases.

Author contributions

All authors contributed to the article; JY: Formal analysis, Conceptualization, Software. HZ Wrote the manuscript. CT: Supervision and revising the manuscript.

Funding

This work was supported by no external funding.

Data availability

The data used in the present study are all publicly available at https://naccdata.org/requesting-data/nacc-data

Declarations

Ethics approval and consent to participate

Not applicable. All data were downloaded from the internet.

Human ethics and consent to participate declarations

Not applicable.

Participate declaration

Not applicable.

Consent for publication

Not applicable.

Clinical trial number:

Not applicable.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.