Associations between disturbing dreams and incident dementia and Alzheimer's disease across diverse international cohorts from the COSMIC collaboration

Darren M. Lipnicki, Ashleigh Simone Vella, Erico Costa, Karen Ritchie, E. Rolandi, Ki Woong Kim, Ding Ding, Yen‐Ching Chen, Jen‐Hau Chen, M SaxtonAlzheimer s & Dementia 2025 · DOI

TL;DR

Disturbing dreams, like nightmares, may serve as an early warning sign for dementia and Alzheimer's disease (AD), particularly in younger older adults (60-69 years) and men. This large international study suggests that tracking dream disturbances could offer a novel, non-invasive way to identify individuals at higher risk for these neurodegenerative conditions.

What Problem Does This Paper Address?

The paper investigates whether disturbing dreams can predict the future development of dementia and Alzheimer's disease (AD). Previous research, primarily in older white American men, suggested a link, but this study aims to broaden that understanding by examining a more diverse international population across a wider age range (60-89 years). The challenge is to determine if this association holds true across different ethnicities and age groups, and if it predicts all-cause dementia or specifically AD. Understanding this potential link is crucial for developing early detection methods and potentially intervening before significant cognitive decline occurs.

Methods

The researchers utilized data from six longitudinal cohort studies across Brazil, China, France, Italy, South Korea, and Taiwan, involving 8,339 participants who were free of dementia or Parkinson's disease at the start of the study. They harmonized the "disturbing dreams" data into a binary (yes/no) format and, in some studies, also captured the frequency of these dreams (none, less than once a week, or once or more per week). The primary analytical approach involved Cox regressions with a random effect for study to account for variations between cohorts. These analyses were stratified by sex and baseline age groups (60-69, 70-79, and 80-89 years). To ensure robustness, analyses were repeated using subsets of studies that included additional covariates such as sleep problems, medications, depression, and APOE*4 status, as well as anxiety.

Datasets

The study integrated data from six existing longitudinal cohort studies:

  • Bambui Cohort Study of Ageing (Brazil)
  • Enquête de Santé Psychologique-Risques, Incidence et Traitement (France)
  • Invecchiamento Cerebrale in Abbiategrasso (Italy)
  • Korean Longitudinal Study on Cognitive Aging and Dementia (South Korea)
  • Shanghai Aging Study (China)
  • Taiwan Initiatives for Geriatric Epidemiological Research (Taiwan)

These datasets provided information on participants' baseline health, cognitive status, and dream experiences, along with a comprehensive set of covariates. No new datasets were introduced.

Key Insights and Ideas

  • Dreams as a potential early biomarker: The findings suggest that changes in dream content, specifically an increase in disturbing dreams, might precede the clinical manifestation of dementia and AD. This opens up the possibility of using dream reports as a non-invasive, accessible, and low-cost preclinical indicator.
  • Age-specific vulnerability: The association between disturbing dreams and dementia/AD risk appears to be strongest in younger older adults (60-69 years), hinting that dream disturbances might be a more sensitive marker in the earlier stages of neurodegeneration.
  • Sex-specific patterns: While disturbing dreams didn't predict dementia in women, they did show a predictive link for AD in men, suggesting potential sex-based differences in how neurodegenerative processes manifest in sleep and dreams.
  • International generalizability: The study's strength lies in its diverse international cohorts, indicating that the observed associations are not confined to specific cultural or genetic backgrounds.

Key Results

  • Disturbing dreams were reported by 24.3% of the overall study population.
  • Overall, disturbing dreams did not predict incident dementia or AD across all age groups and sexes.
  • However, disturbing dreams were associated with a higher risk of dementia in individuals aged 60-69 years across all studies (OR=1.89) and in fully adjusted models (OR=3.93).
  • For AD, disturbing dreams predicted incidence in men aged 60-89 years in fully adjusted models (OR=2.03 for any disturbing dreams, OR=4.00 for frequent disturbing dreams).
  • Disturbing dreams occurring less than once a week (<1/week) showed a significant prediction for dementia (OR=6.41) and AD (OR=4.94) in the 60-69 age group, even in fully adjusted models.

Limitations and Future Work

The authors acknowledge that the definition and reporting of "disturbing dreams" could vary across studies and participants, potentially introducing some subjectivity. While they harmonized the data, nuances might still exist. Furthermore, the study focused on incident dementia and AD, and future research could explore the association with other neurodegenerative conditions or specific subtypes of dementia. The authors suggest that disturbing dreams may be a "pre-clinical indicator and possible prodrome of dementia and AD," highlighting the need for further longitudinal studies to confirm these findings and explore the underlying biological mechanisms. Investigating the impact of interventions targeting sleep or dream content on dementia risk could also be a valuable future direction.

Summary updated

Abstract

Disturbing dreams predicted incident dementia and AD among individuals aged 60-69 years, and incident AD among men aged 60-89 years. Disturbing dreams may be a pre-clinical indicator and possible prodrome of dementia and AD.
BibTeX
@inproceedings{lipnicki2025associations,
  author = {Darren M. Lipnicki and Ashleigh Simone Vella and Erico Costa and Karen Ritchie and E. Rolandi and Ki Woong Kim and Ding Ding and Yen‐Ching Chen and Jen‐Hau Chen and M Saxton},
  title = {{Associations between disturbing dreams and incident dementia and Alzheimer's disease across diverse international cohorts from the COSMIC collaboration}},
  booktitle = {Alzheimer s \& Dementia},
  year = {2025},
  doi = {10.1002/alz70857\_096781},
}
Figure