White matter hyperintensity burden, education and cognition in hospitalised patients with cerebral small vessel disease imaging markers: a cross-sectional study of 199 patients

Background In cerebral small vessel disease (CSVD), white matter hyperintensities (WMH) are associated with cognitive impairment. At the Fazekas visual scale, whether periventricular (PV) and deep (DW) components carry distinguishable information, and how much of the association is independent of education, remain unclear. Methods In 199 consecutive patients with CSVD imaging markers (2024–2025), 112 completed the Montreal Cognitive Assessment (MoCA). The primary analysis was a prespecified linear model of MoCA score on Fazekas score, age and education. Secondary analyses comprised component contrasts, competing models with minimum detectable effects (MDE), tests of non-linearity, and domain analyses with false discovery rate (FDR) control under three specifications. Results Fazekas score was independently associated with MoCA (β = −0.75, 95% CI − 1.23 to − 0.28; P = 0.002). PV and DW were each associated alone (− 1.76 and − 1.09) but did not differ ( P = 0.13); the minimum detectable DW effect equalled the PV estimate. Microbleeds and circulating markers showed no independent association. Severe grades scored 3.70 points lower ( P < 0.001), but categorical and linear models did not differ ( P = 0.08). Visuospatial and orientation domains met the FDR threshold in all three specifications; language was inconsistent. Education explained more variance than WMH (R² 0.24 versus 0.05) and added little to discrimination (area under the curve 0.74 to 0.75). Conclusions WMH burden was independently but modestly associated with cognition; component specificity was not resolvable and education was the dominant correlate. Fazekas grading should inform follow-up intensity rather than decide who needs assessment.