Cerebral Small Vessel Disease Score Associated with Brain Hypoperfusion Predicts Cognitive Decline: A Longitudinal Study

Background This study aims to investigate the association between cerebral small vessel disease (CSVD) score and cerebral blood flow (CBF) at baseline in cognitively intact older adults, and explore whether total CSVD burden serves as an imaging marker can predict longitudinal cognitive impairment. Methods MR images acquired from 509 participants with normal cognition were included in the analysis to assess the association between total CSVD burden score and CBF. Imaging protocols included structural scans, pseudo-continuous arterial spin labeling (pCASL) for CBF quantification, and 3D T1-weighted sequences. CSVD burden scores were rated using a validated 5-point scale by assessing white matter hyperintensity, lacune, perivascular space, microbleed. Participants underwent structured telephone cognitive assessments at a mean follow-up of 7.6 ± 0.1 years post-baseline. The differences in CBF between CSVD burden groups were compared using univariate linear models, and logistic regression analysis was conducted to estimate the risk of longitudinal cognitive impairment. The predictive model were evaluated by the receiver operating characteristic (ROC) curve analysis. Results Severe CSVD scores (score > 2) were significantly associated with decreased CBF in widespread cortical regions ( P _adj < 0.05). The participants with higher CSVD score were more susceptible to longitudinal cognitive decline (OR = 2.995, 95% CI = [1.540, 5.825], P  = 0.001, adjusted for age and sex). The CSVD score model offered good predictive ability for cognitive impairment (AUC = 0.808, P  < 0.001) with an optimal cut-off value of grade>2 (specificity = 88.9%). Conclusion Severe total CSVD score, which is associated with cortical hypoperfusion, serves as an imaging marker of predicting longitudinal cognitive decline. This offers a clinically accessible tool for risk stratification and individualized health monitoring in aging populations.