THE ARISING · Longevity & Healthspan
Study of 247,000 Adults Finds Which Dementia Risk Factor Clusters Matter Most
Dementia risk factors rarely travel alone, and a large Dutch study suggests that the combination may matter: in adults over 65, a cluster of loneliness, depression and physical inactivity was linked to the highest rate of dementia.
This was an observational study in people, published in 2026 in the Journal of Alzheimer's Disease. The researchers, led by van Baal with Köhler, Deckers and Janssen, analysed data from 246,968 adults who took part in the 2012 Dutch Public Health Monitor Survey. Of these, 121,120 were in mid-life (ages 45 to 65) and 125,848 were in late life (over 65). Dementia was followed up through 2023. Observational work can show links. It cannot show that changing a risk factor changes dementia risk.
The team looked at ten modifiable risk factors: obesity, hypertension, diabetes, hearing loss, visual loss, physical inactivity, excessive alcohol consumption, smoking, loneliness and depression. According to the paper, these factors often occur together, but how they group has been poorly described. That makes it harder to spot high-risk people and to design targeted prevention.
To find the groupings, the researchers used a statistical method called latent class analysis. It sorts people into clusters based on which risk factors they tend to share.
In mid-life (average age 55.7), the method found four clusters: cardiometabolic, substance use, psychosocial, and lower risk. In late life (average age 73.9), it again found four: cardiometabolic, substance use, psychosocial and physically inactive, and lower risk.
The researchers then used Cox proportional hazards models, a standard way to compare how quickly events occur in different groups, to test late-life clusters against incident dementia. After adjusting for sociodemographic factors, the psychosocial and physically inactive cluster had the highest hazard ratio, 1.23 (95% confidence interval 1.16 to 1.32). A hazard ratio of 1.23 means a rate of dementia about 23% higher than in the comparison group.
The substance use cluster followed, with a hazard ratio of 1.16 (1.09 to 1.24). The cardiometabolic cluster came next at 1.12 (1.07 to 1.18). The differences are modest, and the ranges for the three clusters overlap.
The authors conclude that the findings underscore the importance of considering risk factor clustering when designing targeted interventions to reduce dementia risk across the life course. The study did not test whether tackling any cluster, or tackling one first, lowers dementia risk, so it cannot say which risks to address first.
Readers should treat the ranking as a pattern in one national survey, not a prescription.
What to watch next
Studies that test whether interventions aimed at these clusters, especially the psychosocial and physically inactive one, actually reduce dementia, and whether the same patterns appear in other countries.