Spatial Relation Between White Matter Hyperintensities and Incident Lacunes of Presumed Vascular Origin: A 14-Year Follow-Up Study

Abstract

Background: The underlying mechanisms of incident lacunes regarding their spatial distribution remain largely unknown. We investigated the spatial distribution pattern and MRI predictors of incident lacunes in relation to white matter hyperintensity (WMH) over 14 years follow-up in sporadic small vessel disease.

Methods: Five hundred three participants from the ongoing prospective single-center Radboud University Nijmegen Diffusion Tensor and Magnetic resonance Cohort (RUN DMC) were recruited with baseline assessment in 2006 and follow ups in 2011, 2015, and 2020. Three hundred eighty-two participants who underwent at least 2 available brain MRI scans were included. Incident lacunes were systematically identified, and the spatial relationship between incident lacunes located in subcortical white matter and WMH were determined using a visual rating scale. Adjusted multiple logistic regression and linear mixed-effect regression models were used to assess the association between baseline small vessel disease markers, WMH progression, and incident lacunes. Participants with atrial fibrillation were excluded in multivariable analysis.

Results: Eighty incident lacunes were identified in 43 patients (mean age 66.5±8.2 years, 37.2% women) during a mean follow-up time of 11.2±3.3 years (incidence rate 10.0/1000 person-year). Sixty percent of incident lacunes were in the white matter, of which 48.9% showed no contact with preexisting WMH. Baseline WMH volume (odds ratio=2.5 [95% CI, 1.6-4.2]) predicted incident lacunes after adjustment for age, sex, and vascular risk factors. WMH progression was associated with incident lacunes independent of age, sex, baseline WMH volume, and vascular risk factors (odds ratio, 3.2 [95% CI, 1.5-6.9]). Baseline WMH volume and progression rate were higher in participants with incident lacunes in contact with preexisting WMH. No difference in vascular risk factors was observed regarding location or relation with preexisting WMH.

Conclusions: The 2 different distribution patterns of lacunes regarding their relation to WMH may suggest distinct underlying mechanisms, one of which may be more closely linked to a similar pathophysiology as that of WMH. The longitudinal relation between WMH and lacunes further supports plausible shared mechanisms between the 2 key markers.

Keywords: incident lacunes; magnetic resonance imaging; small vessel disease; spatial distribution; white matter hyperintensities.

Figure 1.

Rating examples. Baseline: the last scan before the appearance of each incident lacune. Follow-up: the scan on which each incident lacune occurred. A, The incident lacune in left centrum semiovale (center of cross) occurred in normal appearing white matter (no contact with preexisting white matter hyperintensity [WMH]) at baseline and connected with WMH at follow-up. B, The incident lacune (center of cross) appeared in brain region showing partial overlap with preexisting WMH. C, An incident lacune (center of cross) appeared within a preexisting WMH.

Figure 2.

Details in spatial distribution of incident lacunes. This Figure shows the proportion of incident lacunes in brain regions. Forty eight incident lacunes were detected in the subcortical white matter, 26 in basal ganglia‚ and 6 in pons.

Figure 3.

Spatial relationship between incident white matter lacunes and white matter hyperintensities. Twenty-three of the 47 incident lacunes detected in the subcortical white matter were rated grade 0 (48.9%), 22 were rated grade 1 (46.8%), and 2 were grade 2 (4.3%) at baseline. Eleven out of 23 lesions (47.8%) rated as grade 0 were presenting grade 1 on follow-up scans. Three out of 22 lesions (13.6%) rated as grade 1 at baseline became fully inside WMH on follow-up scans.