目的：血管周围间隙（Perivascular space，PVS）是脑小动脉、毛细血管、小静脉周围，与脑实质之间的潜在间隙。PVS扩大时可在头部MRI上被发现，一般认为较多量的扩大的PVS是脑小血管病的影像改变之一，目前的研究对PVS是否与脑萎缩相关仍存在争议，本研究在一项社区人群队列研究中，探究基线不同位置的PVS的严重程度是否与基线时不同脑结构的萎缩相关，并探索基线不同位置PVS的严重程度是否与长程随访的脑萎缩进展相关. 方法：纳入北京市顺义区大孙各庄镇5个自然村35岁以上村民，基线纳入1009人，随访纳入610人，收集人口学资料、既往病史并完成MRI。人工评估腔隙、血管周围间隙严重程度，并利用全自动定量分割方法测定基线及随访MRI各脑结构体积、白质高信号体积及皮层厚度。利用一般线性模型分析不同部位、不同严重程度的PVS与基线、随访时各脑定量结构参数的相关性。利用基于表面的组间分析（Surface-Based Group Analysis，SBM）探索不同部位，不同严重程度的PVS与基线皮层厚度及随访时皮层萎缩进展的空间分布关系。结果：在本社区队列人群中，（1）调整年龄、性别后，基底节区、白质区PVS与基线脑实质分数的相关性均不显著。在调整年龄、性别后，更高的PVS基底节区分级与更大的尾状核体积相关（β0.015，p<0.001）；基线时PVS与多个脑叶厚度升高相关，但不同分级PVS的入组者间,皮层体积分数无明显差异；（2）调整年龄、性别、随访时间后，基线时基底节区、白质区PVS均与5年随访时的全脑萎缩、核团萎缩无关。基线时基底节区PVS与更重的枕叶萎缩相关（β -0.393，p=0.030），SBM提示基线时基底节区PVS与多个脑叶更轻的皮层厚度减少相关，而基线时白质区PVS与各脑叶萎缩无关。（3）不同的年龄段中，PVS严重程度与脑萎缩的关系不同。在较年轻（49岁以下）的人群中，随PVS基底节区严重程度增加，各脑叶的基线皮层厚度呈上升趋势，5年随访时各皮层萎缩进展程度呈降低趋势。在年龄较高的分组（基底节区PVS：61岁以上，白质区PVS：55-61岁）中，随基底节区、白质区PVS分级升高，随访时各皮层及全脑体积萎缩呈加重趋势。结论：总体而言，血管周围间隙与脑萎缩无关。基底节区、白质区PVS的严重程度与基线时脑体积无关，与皮层厚度增高相关。基线时基底节区PVS与更重的枕叶萎缩相关，与多个脑叶更轻的皮层厚度减少相关。本研究未发现基线时白质区PVS与5年随访时的脑萎缩进展的相关性。我们的研究支持PVS在发展的不同阶段对脑萎缩的影响不同,仍需进一步研究。

Background: Perivascular space (PVS) is the potential space between cerebral arterioles, capillaries, venules and brain parenchyma. When PVS is enlarged, it can be found on MRI. It is generally believed that a large amount of enlarged PVS is one of the imaging changes of cerebral small vessel disease. The current study is still controversial on whether PVS is related to brain atrophy. In this community-dwelling cohort study, we explore whether PVS at different locations was associated with atrophy of different brain structures, and explore whether the PVS at different locations at baseline was associated with the progression of brain atrophy at long-term follow-up. Methods: Villagers over 35 years old in 5 natural villages in Shunyi District, Beijing were enrolled. 1009 people were included at baseline and 610 people were included at follow-up. Demographic data, medical history and neuroimaging data were collected. The severity of lacune and perivascular space was manually assessed. the volume of each brain structure, white matter hyperintensity volume and cortical thickness were measured by automatic quantitative segmentation method using baseline and follow-up MRI images. The general linear model was used to analyze the correlation between PVS and brain atrophy at baseline and follow-up. Surface-Based Group Analysis (SBM) was used to explore the relationship between PVS and baseline cortical thickness and the relationship between baseline PVS and cortical atrophy progression at 5-year follow-up was estimatied. Result: In this community cohort population, (1) after adjusting for age and gender, the correlation between PVS and brain parenchyma fraction was not significant. After adjusting for age and gender, severe basal ganglia PVS was significantly associated with greater caudate nucleus volume (β 0.015, p<0.001); PVS was associated with increased cortical thickness, but no significant difference was found in the cortical volume fraction among the participants with different PVS degrees; (2) After adjusting for age, gender, and follow-up time, PVS were not associated with the global brain atrophy and nuclei atrophy at the 5-year follow-up. Baseline basal ganglia PVS was associated with greater occipital lobe atrophy (β -0.393, p=0.030), SBM suggested that those participants with baseline basal ganglia PVS performed lighter cortical thickness reductions, while baseline white matter PVS was not associated with the progression of atrophy. (3) The relationship between the severity of PVS and brain atrophy was different in different age groups. In the younger population (≤ 49 years old）, baseline cortical thickness in each lobe increased with increasing PVS basal ganglia grade, and cortical atrophy progression decreased in each lobe at 5-year follow-up. In the older population（PVS-BG：≥ 61 years old；PVS-WM： 55-61 years old）, with the increase of PVS grade in the basal ganglia, the atrophy of each cortex and the whole brain showed a trend of aggravation during follow-up. Conclusions: In summary, PVS was not associated with brain atrophy. The severity of PVS is associated with increased cortical thickness at baseline, but not with brain volume. Baseline basal ganglia PVS was associated with greater occipital atrophy and less cortical thickness reduction. This study did not find an association between white matter PVS at baseline and progression of brain atrophy at 5-year follow-up. Our study supports that PVS has different effects on brain atrophy at different stages of PVS development, this hypothesis still needs to be further tested.