Cerebral microbleeds (CMB) are key neuroimaging markers of cerebral small vessel disease (CSVD), appearing as round or ovoid hypointense lesions (2–10 mm in diameter) on susceptibility-weighted imaging (SWI; Greenberg et al., 2009). Their presence and burden have been associated with various neurological outcomes, including cognitive impairment, ischemic stroke, and intracerebral hemorrhage (Yin et al., 2025; Lee et al., 2018; Wang et al., 2014). Despite their clinical significance, the pathophysiological mechanisms driving CMB formation remain incompletely understood. Previous investigations have predominantly focused on arterial etiologies, such as hypertensive vasculopathy and cerebral amyloid angiopathy, in explaining the development of CMB (Ungvari et al., 2017; Puy et al., 2024). In contrast, the contribution of venous abnormalities—particularly those involving the deep medullary veins (DMV)—has received limited attention.

Anatomically, DMV radiate from the periventricular white matter toward the subependymal veins, serving a critical role in venous drainage and cerebral homeostasis. Recent advances in SWI have facilitated more detailed visualization of venous structures, prompting increased research interest in DMV involvement in CSVD. Accumulating evidence suggests that DMV dysfunction is closely associated with the severity of CSVD (Tian et al., 2023; Xu et al., 2025). Proposed mechanisms include venous collagen deposition, resulting in venous hypertension and impaired drainage, particularly in distal territories. These changes may lead to increased extracellular fluid accumulation, contributing to microcirculatory disruption and CSVD progression. Moreover, excessive extracellular fluid may exacerbate blood-brain barrier dysfunction and impair the clearance of neurotoxic metabolites (Kern et al., 2024; Li et al., 2024), thereby promoting microstructural damage and potentially facilitating CMB formation (Hayden, 2024). On the basis of this emerging evidence, we hypothesized that DMV dysfunction may contribute to CMB pathogenesis, potentially mediated by pathological increases in extracellular fluid volume.

To test this hypothesis, the present study employed free water (FW) metrics derived from diffusion tensor imaging (DTI) to quantify extracellular fluid content and DMV visibility scores obtained from SWI to assess DMV function. The aim of this study was to examine the interrelationships between DMV scores, FW values, and CMB burden to elucidate potential venous-driven mechanisms underlying CMB development in CSVD.

A total of 176 participants were included in the final analysis. The mean age was 62 ± 12 years, and 81 participants (46.0%) were male. Hypertension was present in 101 individuals (57.4%). CMB were detected in 69 participants (39.2%). The median DMV score was 4 (IQR: 1, 10), and the mean FW value was 0.25 ± 0.02 (Table 1).

This study demonstrated that patients with CMB exhibited significantly higher DMV scores compared to those with mild or absent CMB. Furthermore, DMV scores were independently associated with CMB burden, and this relationship was partially mediated by extracellular FW. The mediating effect of FW remained statistically significant after adjustment for age, hypertension, sex, diabetes, smoking, and hyperlipidemia, suggesting that DMV dysfunction may contribute to CMB pathogenesis through mechanisms involving extracellular fluid accumulation.

As a key component of the cerebral venous drainage system, the DMV plays a critical role in maintaining cerebral hemodynamic balance. Histopathologically, DMV walls consist solely of a single layer of endothelial cells and delicate connective tissue, lacking smooth muscle and elastic fibers (Lahna et al., 2022). Moreover, the DMV has small diameters and long courses with slow blood flow, making them prone to stenosis due to venous collagen deposition. Additionally, the DMV has few anastomotic branches, resulting in limited compensatory capacity once stenosis or occlusion occurs. This anatomical configuration confers limited autoregulatory capacity, making the DMV particularly vulnerable to hemodynamic disturbances. Such vulnerability may predispose the DMV to dysfunction, which can disrupt the cerebral microenvironment.

Progressive collagen deposition in the DMV, often driven by aging and cardiometabolic conditions such as hypertension, dyslipidemia, and diabetes, leads to venous narrowing and luminal stenosis (Moody et al., 1995). Chronic stenosis elevates venous pressure and promotes blood-brain barrier disruption (Keith et al., 2017)—a pathological feature closely linked to CMB formation in preclinical studies (Fisher et al., 2010). Elevated venous pressure also increases vascular permeability, resulting in extravasation of fluid and a rise in extracellular fluid volume. Impaired venous outflow from distal microvasculature may lead to interstitial edema in the cerebral white matter, further contributing to extracellular fluid accumulation (Lan et al., 2024). Excess extracellular fluid impairs the clearance of β-amyloid, facilitating its pathological accumulation. Deposited β-amyloid compromises arterial smooth muscle integrity, enabling the leakage of intravascular components—such as erythrocytes, plasma, and hemoglobin—into the interstitial space (Stokum et al., 2021). The breakdown of these components results in hemosiderin deposition, which appears as hypointense signals on SWI, characteristic of CMB. Additionally, extracellular fluid accumulation can damage axonal and myelin structures, eliciting neuroinflammatory responses that further contribute to CMB development (Sumbria et al., 2016). Neuroinflammation exacerbates blood-brain barrier dysfunction and promotes microvascular hyperpermeability and erythrocyte extravasation (Wang et al., 2021), thereby establishing a vicious cycle. Importantly, neuroinflammation itself may elevate extracellular fluid levels, reinforcing this pathogenic feedback loop.

In this study, DMV scores demonstrated a moderate positive correlation with CMB burden (r = 0.460). On one hand, this suggests complex pathological mechanisms underlying CMB, where age, sex, and hypertension are established high-risk factors for its onset and progression of CMB (Ungvari et al., 2017; Elmståhl et al., 2019; Fandler-Höfler et al., 2024). On the other hand, it indicates interindividual variation in both the degree of DMV stenosis and compensatory capacity. So, in the present study, the mediating role of extracellular FW in the relationship between DMV scores and CMB remained statistically significant even after adjusting for age, sex, vascular risk factors (hypertension, diabetes, smoking, and hyperlipidemia). This finding not only reinforces the robustness of our results but also substantiates a mechanistic link between DMV dysfunction, increased FW content, and CMB burden.

Moreover, our study revealed a significant direct effect between DMV scores and CMB. These findings not only confirm the pivotal role of DMV dysfunction in the pathogenesis of CMB, but also suggest the potential involvement of other unmeasured mediating pathways (e.g., blood-brain barrier disruption or neuroinflammation) in CMB development.

This study has several limitations. First, it was conducted at a single center with a relatively modest sample size, which may introduce selection bias. Future research should incorporate multi-center cohorts and larger sample sizes to enhance generalizability. Second, the cross-sectional design limits the ability to assess temporal relationships between DMV impairment, FW elevation, and CMB development. Longitudinal studies are needed to elucidate causality. Third, clinical cognitive data were not collected, preventing direct evaluation of whether elevated FW contributes to cognitive decline in individuals with CMB. Fourth, factors potentially contributing to elevated FW values may include blood-brain barrier dysfunction and neuroinflammation. However, the absence of blood-brain barrier function and neuroinflammation data in our study precluded their inclusion as confounding factors for statistical adjustment.

In conclusion, the dysfunction of the DMV partially mediates the formation of CMB in patients with CSVD through increased extracellular FW. These findings suggest that impaired venous drainage and interstitial fluid accumulation may constitute key pathological processes in the development of CMB.