INTRODUCTION

The prognostic utility of plasma proteomics for cognitive decline and dementia in a Southeast Asian population characterized by high cerebrovascular disease (CeVD) burden is underexplored.

METHODS

We examined this in a Singaporean memory clinic cohort of 528 subjects (n = 300, CeVD; n = 167, incident cognitive decline) followed-up for 4 years.

RESULTS

Of 1441 plasma proteins surveyed, a 12-protein signature significantly predicted cognitive decline (q-value < .05). Sixteen diverse biological processes were implicated in cognitive decline. Ten proteins independently predicted incident dementia (q-value < .05). A unified prediction model combining plasma proteins with clinical risk factors increased the area under the curve for outcome prediction from 0.62 to 0.85. External validation in the cerebrospinal fluid proteome of an independent Caucasian cohort replicated four of the significantly predictive plasma markers for cognitive decline namely: GFAP, NEFL, AREG, and PPY.

DISCUSSION

The prognostic proteins prioritized in our study provide robust signals in two different biological matrices, representing potential mechanistic targets for dementia and cognitive decline.

HIGHLIGHTS

A total of 1441 plasma proteins were profiled in a Singaporean memory clinic cohort. We report prognostic plasma protein signatures for cognitive decline and dementia. External validation was performed in the cerebrospinal fluid proteome of a Caucasian cohort. A concordant proteomic signature was identified across both biofluids and cohorts. Further studies are needed to explore the therapeutic implications of these proteins for dementia.