OBJECTIVE

To explore the effect of brain atrophy on the cognition in patients with subcortical ischemic vascular disease (SIVD).

METHODS

A total of 116 SIVD patients were enrolled from the Department of Neurology, First Affiliated Hospital, Anhui Medical University between September 2013 and December 2014. Lobar atrophy, leukoaraiosis (LA), lacunar infarcts (LI) and vascular risk factors were analyzed in the 116 SIVD patients who were divided into three groups according to the diagnostic criteria: non-cognitive impairment group (SIVD-NCI) , mild cognitive impairment group (SIVD-MCI) and dement group (SIVD-VaD). All patients underwent magnetic resonance imaging (MRI) with a 3.0-T system. The cognitive functions were evaluated by mini-metal state examination (MMSE), the Cambridge cognitive examination-Chinese version (CAMCOG-C), etc. A widely used visual atrophy rating method (0 to 3) was adopted to rate the severity of frontal, parietal and temporal lobe atrophy. The degree of LA and the numbers of LI in 4 brain regions (frontal, parieto-occipital, temporal, and basal ganglia) were evaluated meanwhile.

RESULTS

Firstly, both the SIVD-MCI and SIVD-VaD groups showed significantly higher total scores of atrophy, higher frontal lobe atrophy scores, higher LA scores and larger LI numbers than SIVD-NCI (H=6.138, P=0.013; H=45.845, P=0.000; H=36.818, P=0.000; H=37.46, P=0.000). There were no significant differences in temporal lobe atrophy scores between SIVD-NCI group and SIVD-MCI group. Parietal lobe atrophy scores also showed no differences among the three groups. Secondly, as well as total numbers of LI, total scores of atrophy and LA were negatively correlated with SIVD cognition, especially frontal lobe atrophy scores, parieto-occipital LA scores and basal ganglia LI numbers had a remarkable negative correlation with MMSE scores, CAMCOG-C scores and partial subitems in CAMCOG-C scores (P<0.005). However temporal LI numbers was absence of correlation with MMSE scores, CAMCOG-C scores and subitems in CAMCOG-C scores (P>0.005). Thirdly, education (OR= 0.787, P=0.009), total scores of LA (OR= 1.201, P=0.036) and total numbers of LI (OR= 1.221, P=0.011) remained to be associated with cognition after controlling for sex, age, education, hypertension, diabetes, smoking. While, after adjusting atrophy scores, total scores of LA had no significant association with cognition. Atrophy scores, total numbers of LI and education still showed significant association with cognition, particularly frontal lobe atrophy scores (OR= 16.082, P=0.000), but not temporal and parietal lobe atrophy scores.

CONCLUSIONS

Brain atrophy may be a new and independent predictive index of cognitive impairment in SIVD. Further, the effect of brain atrophy, the numbers of LI and degree of LA on cognitive decline is independent and decreases in turn.