Multimarker Cerebral Small Vessel Disease Score and Risk of Incident Dementia in the Framingham Heart Study.

BACKGROUND AND OBJECTIVES

Individual MRI markers of cerebral small vessel disease (CSVD) are associated with impaired cognition and dementia but may not reflect the overall burden of CSVD. In addition, it is unclear whether these markers provide additional value in dementia risk assessment beyond vascular risk factors alone. Thus, we studied the association between the additive burden of multiple CSVD markers and incident dementia and determined whether this relationship remains independent of the Framingham Stroke Risk Profile (FSRP), a tool used commonly used for stroke risk prediction.

METHODS

A total of 1,152 MRI scans from participants in the Original and Offspring cohorts of the Framingham Heart Study, a large observational cohort study, were included. Participants were older than 55 years and free of prevalent dementia, stroke, or other neurologic conditions at the time of MRI. A multimarker score capturing CSVD burden was defined as the sum of CSVD features detected in the MRI: cerebral microbleeds, covert brain infarcts, extensive white matter hyperintensities, high-burden perivascular spaces, and cortical superficial siderosis. Multivariate Cox regression models examined the association between the multimarker CSVD score and incident all-cause dementia, Alzheimer dementia (AD), and vascular dementia.

RESULTS

The mean age was 70.9 years (SD 8.7) (527 [46%] were male), and 211 (18%) had a CSVD score of ≥2. Over a median follow-up time of 7.4 years (interquartile range 4.6-11.3), participants with a score ≥2 had significantly elevated risk of all-cause dementia compared with those with no CSVD markers after adjustment for the FSRP (hazard ratio [HR] 1.67; 95% CI 1.05-2.66) and vascular risk factors (HR 1.76; 95% CI 1.10-2.81). The multimarker CSVD score demonstrated similar model performance metrics to the FSRP (Harrell -statistics 0.82-0.83).

DISCUSSION

We found a significant association between all-cause dementia and multimarker CSVD scores, which was independent of the FSRP as well as its individual components. Our results support the use of a multimarker CSVD score as an indicator for incident all-cause dementia risk and suggest that it may be as robust as the FSRP. Further studies are necessary to validate the use of a multimarker CSVD score in dementia risk prediction.