Age differences in periventricular and deep white matter lesions

Abstract

Deep white matter hyperintensity (DWMH) and periventricular (PV) white matter lesion volumes are associated with age and subsequent stroke. We studied age differences in these volumes accounting for collinearity and risk factors. Subjects were 563 healthy family members of early-onset coronary artery disease patients. Using 3T magnetic resonance imaging, lesions were classified as DWMH or PV. Age association with lesion classification was analyzed using random effects Tobit regression, adjusting for intracranial volume (ICV) and risk factors. Subjects were 60% women, 36% African-American, mean age 51 ± 11 years. In multivariable analysis adjusted for PV and ICV, DWMH was associated with age (p < 0.001) and female sex (p = 0.003). PV, adjusted for DWMH and ICV, was age associated (p < 0.001). For each age decade, DWMH showed 0.07 log units/decade greater volume (95% CI = 0.04-0.11); PV was 0.18 log units/decade greater (95% CI = 0.14-0.23); slope differences (p < 0.001). In people with a family history of coronary artery disease, PV and DWMH are independently and differentially associated with age controlling for traditional risk factors.

Keywords: Coronary; Imaging; Risk factors; White matter disease; Women and minorities.

Figure 1

(a) and (b). Axial magnetic resonance imaging showing the position of nearly exclusive perventricular white matter lesions (a) and deep white matter lesions (b)

Figure 2. Scatter plot of DWMH and PV vs. age

Deep white matter hyperintensity volumes (filled circles and solid line) and periventricular hyperintensity volumes (hollow symbols and dotted line) on the logarithmic scale vs. age and best fit unadjusted linear fit. Volumes below the level of detection have been plotted at the level of half the minimum detectable volume. Small random numbers have been added to the plotted points so that overlaid symbols can be seen separately. The unadjusted slope for DWMH (7% higher volume per year) was lower than that for PV (12% higher volume per year, p < 0.001)