History of falls is associated with increased risk of future dementia in adults aged 40 years and older.

This systematic review and meta-analysis investigated the relationship between a history of falls and the subsequent development of dementia. The study population comprised 2,922,624 middle-aged and older adults aged 40 years and older who did not have dementia at baseline. The specific setting of the included studies was not reported. The primary exposure was a history of falls, categorized as either a single fall or multiple falls. The primary outcome measured was the future development of all-cause dementia. No secondary outcomes were reported in the available data. The analysis included five studies contributing to the pooled estimates.

The results demonstrated a clear association between fall history and dementia risk. For the pooled incidence of future dementia in adults aged 40 years and older with a history of falls, the rate was 11.6% (95% CI, 4.2%-19.0%). In the subgroup of older adults aged 60 years and older, the pooled incidence was 12.3% (95% CI, 4.7%-20.0%). When analyzing specific fall patterns, a history of a single fall was associated with an adjusted hazard ratio (aHR) of 1.20 (95% CI, 1.07-1.36) for future all-cause dementia. A history of multiple falls showed a stronger association, with an aHR of 1.74 (95% CI, 1.53-1.98).

Safety and tolerability data were not reported in the included studies, as the research focused on observational associations rather than an intervention with adverse event monitoring. Consequently, no information regarding adverse events, serious adverse events, discontinuations, or tolerability is available for this analysis. The study design inherently limits the ability to assess safety profiles of the exposure itself.

These findings must be contextualized within the existing literature on dementia risk factors. While prior landmark studies have identified vascular risk factors and genetic markers, this analysis highlights falls as a potential clinical marker. The suggested dose-response relationship, where multiple falls confer a higher risk than a single fall, aligns with the clinical observation that recurrent falls may indicate underlying neurological or physiological vulnerability. However, the evidence base remains limited to five studies, which restricts the generalizability of these specific estimates.

Several significant methodological limitations must be acknowledged. The meta-analysis reported very high heterogeneity, with an I-squared value of 99.8%. This indicates substantial variability in the results across the included studies, which may stem from differences in study populations, definitions of falls, or dementia diagnostic criteria. The wide confidence intervals for the pooled incidence rates further reflect this uncertainty. Additionally, the limited number of studies included in the meta-analysis reduces the statistical power and robustness of the conclusions.

Clinically, these results suggest that recurrent falls may serve as a potential marker for identifying individuals at higher risk of dementia. Clinicians should maintain heightened vigilance for cognitive decline in middle-aged and older adults with a history of recurrent falls to facilitate early detection. However, because the evidence is observational, falls are a predictor rather than a cause of dementia. The association does not imply that preventing falls will necessarily prevent dementia, though addressing fall risk remains a standard of care.

Important questions remain unanswered regarding the mechanisms linking falls to dementia. It is unclear whether falls are a prodromal symptom of neurodegenerative disease or if they result from shared risk factors such as frailty or vascular disease. Further high-quality research is warranted to clarify this association, ideally through larger prospective cohorts that can better control for confounding variables and reduce heterogeneity. Until then, the findings should be interpreted with caution, recognizing that the current evidence base is limited and the certainty of the association is low.