2025 Volume 19 Issue 1 Article ID: ra.2025-0020
Over the past decade, clinicians and researchers have increasingly recognized the significance of the glymphatic system. Evidence demonstrates that this system—named for its reliance on astrocyte endfeet of glial cells and its lymphatic-like waste clearance function from the brain—is essential for regulating the accumulation and removal of amyloid aggregates and other interstitial waste products that may cause cognitive decline if not removed. Its activity is highly regulated, with flow driven by arterial wall pulsatility linked to the cardiac cycle, facilitating perivascular cerebrospinal fluid (CSF) influx into the brain interstitium and its efflux into the venous system. In the present review, we highlight the interplay between the glymphatic system and neurovascular diseases, as well as conditions that are currently being treated by endovascular means, including subarachnoid hemorrhage, idiopathic intracranial hypertension, steno-occlusive disease, and arteriovenous shunting diseases. We describe how changes in arterial pulsatility, disturbances in para-arterial CSF influx, changes in aquaporin-4 receptor composition, or venous hypertension with a decreased arteriovenous pressure gradient can cause dysfunction of different components of the glymphatic system, leading to similar clinical symptomatology with progressive cognitive decline that may be reversible.
In a series of 4 studies published between 2012 and 2013, the group of Maiken Nedergaard described the anatomical organization, function, and physiological regulation of a novel system responsible for maintaining homeostasis within the central nervous system (CNS), including nutrient delivery, waste removal, and consistency of the ionic microenvironment.1–4) They coined this system the glymphatic system (GS), given its dependence on glial cells and its similarities to the lymphatic system. As a cerebral drainage pathway, this system led to a fundamental change in our understanding of the circulation of fluids, including the cerebrospinal fluid (CSF) and the interstitial fluid (ISF), within the CNS.5) In brief, and somewhat oversimplified, CSF egresses through the perivascular spaces (PVSs) from the subarachnoid space into the brain’s interstitial space. This egression is modulated by aquaporin-4 (AQP4) channels in astrocytes surrounding the PVSs.6) From the brain interstitium, where the CSF “bathes” the brain tissues, the fluid (and the waste products it has collected through its journey through the neuropil) is then drained into the perivenous spaces, reaches the meningeal lymphatic (ML) system, and, finally, the deep cervical lymph nodes.7) In this “convective flow” model, influx of CSF into the brain interstitium is balanced by perivenous efflux of ISF. While the GS and disorders thereof have been put forward in a variety of neurological conditions, including Alzheimer’s disease, Parkinson’s disease, and other neurodegenerative disorders,8) this system may, in a broader sense, also be altered by various neurovascular diseases, some of which can be modulated through endovascular techniques, a field of research that has only recently been explored in greater detail.9)
In this review, we will first briefly describe the anatomy of brain compartments and fluid pathways as well as its barriers, and then focus on the physiology of the neurovascular unit (NVU) that comprises the GS. Subsequently, we will highlight 4 major neurovascular disease groups (subarachnoid hemorrhage [SAH], idiopathic intracranial hypertension [IIH], arterial steno-occlusive disease, and arteriovenous shunting diseases), where we believe the GS plays a role in disease manifestation and where endovascular techniques may play a role in modulating glymphatic structure, flow, and function, thus deepening our understanding of these diseases and their potential management.
In order to understand the anatomy and physiology of the GS, it is helpful to review the basics about the various fluids in the CSF, their compartments, and their interactions with each other: intracranial extracellular fluid exists as plasma, ISF, and CSF. Plasma contains significantly more protein, sugar, and amino acids than ISF and CSF, while potassium and calcium ion concentrations in ISF and CSF are about half those in plasma.10) These differences are maintained by physical barriers like tight junctions in the blood–brain barrier (BBB) and the blood–CSF barrier. Mitochondria-rich endothelial and ependymal cells, equipped with ionic and molecular transporters, create osmotic gradients essential for CSF and ISF production.11)
ISF (280–300 mL in adults) is about twice the volume of CSF (140–150 mL) and serves different functions.10) ISF facilitates nutrient and oxygen delivery and waste removal for neurons and glia, while CSF cushions the brain, provides buoyancy, and helps regulate solute transport through bulk flow driven by cardiac and respiratory cycles. Approximately 90% of CSF is produced by the choroid plexus, with daily production ranging from 450 to 500 mL, regulated by circadian hormonal changes and requiring continuous efflux.12)
CSF drains through multiple pathways, including meningeal and nasal lymphatics, dural nerve sleeves, and arachnoid granulations or villi.13) Exchange between CSF and ISF occurs along the ependyma of ventricles and PVSs surrounding vessels from the subarachnoid space. Understanding this exchange may offer insights into brain homeostasis and pathologies linked to its disruption.14)
As superficial arteries and veins penetrate the cerebral cortex from the subarachnoid space, they are enclosed by CSF-filled channels within a leptomeningeal sheath. The PVS around penetrating arteries lies between the endothelial basement membrane and the glia limitans, enabling fluid transport and communication with the basal lamina of capillaries. Animal studies show ~1 μm pores in the leptomeningeal and pia mater layers, allowing CSF movement from the subarachnoid to the subpial space.15)
The glia limitans, formed by astrocytic endfeet, covers the majority of the cerebral vasculature and contains AQP4 channels highly polarized on the luminal surface, aiding water transport.16) These channels facilitate fluid movement between the perivascular and interstitial spaces. An important prerequisite for understanding CSF fluid flow is that the perivascular influx of CSF into the brain is directional—with CSF entering the brain exclusively within the periarterial spaces and ISF leaving the brain through perivenous channels.10) Due to its dependence on the glial AQP4 channel and its pseudo-lymphatic function, Nedergaard’s group named this pathway of periarterial CSF inflow and perivenous ISF and solute drainage the glial-associated lymphatic pathway, or glymphatic pathway.1)
Water from the paravascular CSF can enter the interstitial space via 2 distinct periarterial pathways: through approximately 40-nm gaps in astrocytic endfoot processes, which act as size-selective filters for molecular tracers, and via AQP4 water channels in the astrocytic endfeet.12) Once in the interstitial space, this water is believed to facilitate convective transport of solutes and waste products, such as neurotoxic amyloid β (Aβ) oligomers, into the perivenous CSF space, where they are subsequently removed through CSF efflux mechanisms.1) The CSF influx is believed to be controlled by penetrating arterial pulsatility, which integrates changes related to amplitude and frequency of PVS diameter oscillation.17) However, arterial pulsatility is not the only parameter modulating the glymphatic system GS. Increased glymphatic function during sleep has been associated with a decline in central norepinephrine levels, leading to expansion of the extracellular space, and a resultant decrease in tissue resistance, which in turn leads to faster CSF influx and interstitial solute efflux.4) In addition, venous pressure will also modulates the GS: due to the compliance of cerebral veins, which behave as distensible or collapsible channels depending on transmural pressure (i.e., the difference between the surrounding CSF pressure and the intraluminal pressure), less CSF will be reabsorbed through perivenous channels in case of venous hypertension,18) a pathological mechanism that will be discussed in greater detail in subsequent sections.
While our understanding of the GS is ever evolving, it becomes clear that this system relies on 3 key elements: an intact BBB, which facilitates unidirectional ionic influx into the PVS followed by water; the PVS, which acts as a water reservoir for the interstitium and a transport pathway for solutes toward CSF efflux routes; and astrocytic endfeet of the glia limitans, which regulate solute exchange rates and size selectivity, as well as water influx, based on their size and degree of vascular ensheathment, all of which lead to CSF egressing into the neuropil and the transport of fluids through the interstitium toward the venous system, where it is reabsorbed into the perivenous CSF spaces.19)
The GS plays a vital role in neurophysiology, particularly in waste clearance. In AQP4 knockout mice with reduced glymphatic function, the clearance of interstitial solutes like mannitol and Aβ is significantly impaired.1) Enhanced glymphatic activity is also linked to reduced brain lactate levels during the transition from wakefulness to sleep. Inhibition of glymphatic clearance in anesthetized mice—via AQP4 deletion, acetazolamide treatment, cisterna magna puncture, or altered head position—results in elevated brain lactate levels and reduced lactate in cervical lymph nodes. Beyond waste removal, the glymphatic pathway is essential for distributing nutrients, such as glucose, across the brain19) (Fig. 1).
However, it has been the clearance function of the GS that has sparked the most interest in understanding the GS under normal and pathological circumstances in recent years, as decreased clearance with subsequent accumulation of pathological solutes, particularly amyloid, with Aβ plaque formation, has been linked to long-term neurodegeneration, cognitive decline, and Alzheimer’s disease.12)
Various noninvasive imaging techniques to visualize the GS have been proposed; of these, the diffusion tensor image analysis along the PVS technique appears to be the most promising given its noninvasive nature. This technique measures water diffusivity along the PVS and compares this diffusivity to the diffusivity along normal projection and association fibers.20) In a healthy GS the water diffusion should be high along the PVS as a measure of glymphatic outflow. The technique has been used in studies on Alzheimer’s disease,20) epilepsy,21) and IIH.22) Although this tool appears promising, it has become evident that a single method alone will not fully capture the complexity of the GS. Therefore, combining multiple techniques would be beneficial for a more comprehensive assessment.20)
In the following, we will describe various neurovascular diseases and their interplay with the GS and describe how glymphatic dysfunction can occur through perivascular blockage, channelopathies, decreased pulsatility, or decreased arteriovenous pressure gradients that interfere with different components of the GS but may lead to common clinical manifestations (Figs. 2 and 3).
SAH is most commonly caused by rupture of cerebral aneurysms (aneurysmal SAH: aSAH) or trauma. Vascular malformations such as arteriovenous malformations (AVMs), dural arteriovenous fistulas (dAVFs), and cerebral amyloid angiopathy (CAA), as well as cortical or venous sinus thrombosis, can also be potential causes.23) Among these, aSAH is the most common and severe form of SAH and harbors a different natural history, as it is more commonly associated with complications such as early brain injury (EBI) and delayed cerebral ischemia (DCI). EBI occurs within the first 3 days after onset of aSAH and is characterized by transient global ischemia due to increased intracranial pressure (leading to decreased cerebral perfusion pressure), direct brain injury caused by the hemorrhage, microcirculatory disturbances due to blood in the subarachnoid space, and cortical spreading depolarization. These factors contribute to inflammation, oxidative stress, and BBB disruption, ultimately leading to brain edema, neuroinflammation, apoptosis, and neuronal cell death. On the other hand, DCI occurs between days 4 and approximately 14 after onset and is mainly characterized by focal cerebral ischemia due to cerebral vasospasm and the progression of EBI pathophysiology.23) In addition, SAH is known to cause various complications even in the subacute to chronic phases, such as hydrocephalus, persistent cognitive disturbances, and various hormonal abnormalities due to hypothalamic-pituitary axis dysfunction. These complications are thought to result from the residual hematoma, inflammation, and immune substances causing blood and CSF circulation disturbances, including GS dysfunction.23)
The treatment of aSAH aims primarily at preventing rebleeding of the cerebral aneurysm, either through endovascular or open surgical approaches. Prevention and treatment of DCI include Triple-H therapy (hypertension, hemodilution, and hypervolemia), systemic and/or intra-arterial administration of nimodipine for CVS, and in some cases, percutaneous transluminal balloon angioplasty (PTA).23) Despite current treatments, SAH remains the most fatal form of stroke, with a case fatality rate of approximately 40%.23)
The impact of SAH on the GSThe circulation and excretion of CSF and ISF primarily involve 2 key pathways: the GS, which regulates fluid dynamics within the brain parenchyma, and the ML system, which facilitates the clearance of waste products and immune cells from the CSF in the meninges. Additionally, the GS relies on the water channel protein AQP4 to regulate CSF-ISF flow, playing a crucial role in maintaining fluid balance.24,25) Based on these considerations, the following section describes the impact of SAH on each system.
One of the earliest consequences of aSAH is the dysfunction of the PVS, which serves as a conduit for CSF and ISF exchange. Blood products and inflammatory mediators released during aSAH accumulate in the PVS, reducing CSF influx into the interstitium and thus impeding waste clearance.24–27) Studies suggest that PVS dysfunction is most severe within 6 hours to 3 days post-SAH and can persist long-term, thereby contributing to secondary brain injury.25)
AQP4, a water channel expressed on astrocytic endfeet, plays a crucial role in GS function by facilitating CSF–ISF exchange.24) Following SAH, AQP4 polarization is significantly disrupted, leading to BBB breakdown, impaired waste clearance, and exacerbated brain edema.28) Dysfunctional AQP4 further contributes to neuroinflammation and oxidative stress, aggravating neuronal damage.24,28)
The ML system is responsible for draining excess CSF and immune cells, supporting immune surveillance in the brain. SAH-induced coagulopathy and inflammatory cytokine release impair ML drainage, leading to lymphatic congestion, prolonged immune activation, and further exacerbation of neurological injury.25,29) Animal studies demonstrate that ML dysfunction persists for at least 1 week post-SAH and is associated with Tau protein accumulation, lymphocyte infiltration, and glial activation.30)
Recovery of GS function remains incomplete even 2 weeks post-SAH, with persistent PVS enlargement linked to chronic inflammation, fibrosis, and BBB damage.30) Additionally, post-SAH sleep disturbances—often due to hypothalamic dysfunction—further delay GS recovery and contribute to long-term neurodegeneration.25) Evidence also suggests that prolonged GS dysfunction leads to abnormal accumulation of neurotoxic proteins such as Aβ and Aτ, increasing the risk of cognitive decline and neurodegenerative diseases.31)
Imaging evidence of GS dysfunction after SAHAdvancements in imaging technology have enabled the assessment of GS function and treatment effects in SAH animal models and human studies.24,27,32) Kim et al. analyzed PVS changes in aSAH patients using serial MRI and reported increased visualization of enlarged PVS (centrum semiovale >basal ganglia), where it was independently correlated with a high burden of aSAH.32) Yu et al., on the other hand, found that enlarged PVS (basal ganglia >centrum semiovale) has a strong association with aSAH patients, even without pre-existing vascular risk factors.27) These findings suggest the potential use of non-contrast MRI-based PVS evaluation as a marker of GS dysfunction and a predictor of long-term cognitive impairment in SAH patients.27,32)
GS-targeted therapeutic approaches for aSAHVarious pharmacological and surgical interventions have been proposed to improve GS function after aSAH, primarily based on animal models.
Nimodipine enhances GS function by activating the cAMP/PKA pathway, reducing brain edema, neurological deficits, and AQP4 dysregulation, showing efficacy in both DCI and EBI.33) Tissue plasminogen activator promotes hematoma dissolution, enhances GS clearance, and reduces inflammatory cytokines, mitigating brain edema.32) Tissue factor (TF)-neutralizing antibodies and thrombin inhibitors (argatroban, dabigatran) may counteract the effect of thrombin and TF in causing GS pathway disruption, thus reducing inflammation and BBB dysfunction.34) SAH-induced reactive astrocytosis leads to A1-type astrocyte activation, promoting neurotoxicity and GS dysfunction. PEDF-34, ponesimod, and gastrodin suppress A1 astrocyte activation, reducing inflammation and oxidative stress.28) Additionally, inhibiting the mTOR pathway and protease activity has shown promise in reducing posthemorrhagic fibrosis and improving CSF circulation and GS function.30)
An AQP4-specific inhibitor, TGN-020, has demonstrated the ability to amplify AQP4 polarization and activation of astrocytes subsequent to cerebral ischemia–reperfusion injury and SAH, mitigating glymphatic dysfunction.28) β-Hydroxybutyrate has also been shown to promote AQP4 polarization recovery by upregulating the expression of membrane-associated syntrophin α1, suppress neuroinflammation, and enhance BBB protection, making it a potential candidate for improving GS function during SAH recovery.35)
CSF outflow via the ML system is significantly impaired after SAH, with occlusions observed in the olfactory, optic, facial, and vestibular nerve drainage pathways in animal studies.25) Ventricular or lumbar drainage, as well as endoscopic third ventriculostomy, may restore CSF outflow and facilitate the recovery of both GS and ML systems, improving neurological and cognitive recovery.25)
Endovascular therapies, including PTA, have been reported to enhance cerebral perfusion and alleviate hypoperfusion-related GS dysfunction.23) One could hypothesize that the increased pulsatility that can be obtained following restoration of proximal vessel diameters may play a role in the ameliorated function of the GS. For aSAH cases requiring open surgery, basal cisternostomy has been suggested to lower basal cistern pressure, reverse the intracranial pressure (ICP) gradient, and facilitate GS function and brain edema resolution.36)
IIH is a clinical condition characterized by raised ICP without an identifiable cause. The prevalence of IIH has continued to rise over the past decade, reaching prevalence of approximately 10 individuals per 100000 in 2022.37) It predominantly affects obese females of reproductive age. Headache is the most common clinical presentation, followed by pulsatile tinnitus, visual disturbances, and cognitive decline. The pathophysiology of IIH remains unclear, previously postulated to be related to altered CSF homeostasis, either through increased CSF production via the choroid plexus or impaired CSF absorption through the cerebral venous system (Fig. 4).
The connection between IIH and obesity has been extensively studied. ICP opening pressure, prognosis, and risk of recurrence in IIH are believed to be associated with the degree of centripetal fat distribution rather than body mass index (BMI).38) Obesity is a chronic inflammatory condition linked to various proinflammatory cytokines, chemokines, adipokines, and hormones. These biochemical changes, such as androgen excess, increased leptin levels, and IIB-hydroxysteroid dehydrogenase type 1 (IIB-HSD 1) dysfunction with elevated cortisol, may lead to increased CSF production. Elevated pro-inflammatory cytokines, on the other hand, may induce fibrosis of the arachnoid villi in the dural venous sinuses or astrogliosis, reducing CSF absorption.38,39)
Insulin regulates the expression of AQP4, which is an important component of the NVU, controlling the movement of water and metabolites across the ventricles, brain parenchyma, and perivascular astrocytic endfeet processes. Obesity has a strong association with insulin resistance, which may contribute to dysregulation of AQP4.38) Dysregulation of AQP4 in IIH is evidenced by decreased CSF level of AQP4 and astrocyte hypertrophy (compensatory immunoreactivity) in biopsies of the frontal cortex from IIH patients.39)
CSF clearance pathways and the impact of congested GSThe current working model for brain CSF clearance consists of 3 main components:40) periarterial inflow of fluid from the subarachnoid spaces at the brain surface to the parenchymal interstitium; intraparenchymal flow of fluid through the brain interstitium; and clearance of fluid from the interstitium into the perivenous, perineural, and ML spaces.
Congestion of the GS, or impaired clearance of ISF to the dural venous sinuses, has been postulated to trigger a cascade of events resulting in IIH. Glymphatic congestion can be assessed on imaging, based on the number of PVSs (as part of the glymphatic clearance pathway)41) and diffusivity along the projection and association fibers adjacent to the lateral ventricle as a measure of glymphatic function (using the diffusion-weighted image analysis along the perivascular space [DWI-ALPS] technique).42) Based on an MR imaging assessment study of 36 IIH adults, Jones et al. found a significantly increased number of visible PVSs in IIH, indicative of glymphatic dysfunction.41) A study assessing 81 IIH patients, comparing the ALPS-index (a measure of glymphatic clearance function) of IIH patients to a healthy control group, concluded that IIH patients had significantly lower glymphatic clearance (ALPS-index) compared to the healthy control group,42) further supporting the theory of glymphatic congestion in IIH.
As a compensatory mechanism, arachnoid granulation may overgrow to improve CSF drainage from the GS to the dural venous sinuses; however, this overgrowth of arachnoid granulation may in turn result in intrinsic obstruction of the dural venous sinuses. Glymphatic congestion, if decompensated, may cause raised ICP leading to extrinsic compression of the dural venous sinuses. These processes may result in restriction of venous outflow and thus worsening glymphatic congestion. Bilateral transverse sinus stenosis (TSS) was found in 83%–94% of IIH patients. There is no correlation between the degree of TSS and the clinical course among patients with IIH, suggesting that clinical symptomatology, not the degree of TSS, should be used to determine management in IIH.43)
In the absence of effective venous outflow, CSF will be diverted through the ML system, and this overflow of CSF in the ML system will be shown on imaging as enlarged, tortuous optic nerve sheaths (along with protrusion of the optic nerve discs and flattening of the posterior globes) and an empty sella (herniation of the subarachnoid space through the diaphragmatic sella).40) Decompensation of the ML system will lead to increased ICP and various symptoms of IIH. Timely intervention would be required to prevent disease progression.
Treatment of IIH and future developmentsTreatment of IIH aims to provide symptom reliefe and to prevent visual deterioration. Understanding the pathophysiology of IIH development is important in refining treatment strategies.
The first step in treating IIH is addressing obesity, its primary risk factor. Weight loss in the range of 3%–24% has been reported to lead to remission, and weight regain has been found to be a risk factor for disease recurrence.44) Weight loss therapy consisting of multicomponent lifestyle intervention (diet + physical activity + behavior) has the most robust evidence for weight loss with a BMI <35 kg/m2, while bariatric surgery should be considered for BMI ≥35 kg/m2.44)
Carbonic anhydrase inhibitors such as acetazolamide and topiramate decrease CSF production by inhibiting the Na+/K+ ATPase transporter. Acetazolamide is the most established first-line medical therapy for IIH. Experimental drugs that regulate CSF production, such as the glucagon like peptide-1 receptor agonist, exenatide and the 11β-HSD1 inhibitor AZD4017, are currently still under evaluation for their use in IIH.45)
Surgical options for IIH include CSF diversion procedures such as lumbar puncture (LP) and ventricular shunts, and non-CSF diversion procedures such as optic nerve sheath fenestration (ONSF) and venous sinus stenting (VSS). LP is often the first-line surgical therapy, as it is minimally invasive and provides diagnostic information such as CSF opening pressure and CSF composition, as well as clinical response post-CSF drainage to predict future treatment response with other surgical treatments. Ventriculoperitoneal (VP) shunts are highly effective in relieving IIH symptoms, with over 85% improvement in headaches, papilledema, and visual impairment.46) However, approximately half of these patients with a VP shunt require shunt revision,46) raising the question of its long-term durability. Lumboperitoneal shunt or lumbocaval shunt are other less established alternatives with varying effectiveness depending on institutional experience.
ONSF has been shown to be effective at reducing optic disc swelling (97%), improving visual field (76%), and visual acuity (41%).47) Early institution of treatment is important to restore vision before the onset of optic nerve atrophy.
VSS for IIH is a relatively low-risk procedure, with a 2%–4% reported major complication rate,48) such as intracranial hemorrhages (especially subdural hematoma) from venous perforation, arterial dissection, or stent thrombosis. VSS has a high rate of short-term symptomatic relief, especially for visual symptoms, papilledema, and pulsatile tinnitus, but only 30% of patients achieve long-term symptom resolution. Symptom recurrence rate is as high as 60%, and some experience more intensified or new symptoms compared to their pre-stenting state. Most of these patients improved with regular LPs; however, about 17% in-stent (or adjacent stent) stenosis required additional stenting or CSF shunting procedure.49) Previous studies indicate that post-stenting restenosis can be reversed with CSF shunting procedures, such as LP and ventricular shunt. This suggests that TSS is likely a result of raised ICP in IIH, rather than its cause.50) VSS should therefore be performed only for medically refractory IIH, rather than being routinely applied for all cases of IIH.
Current surgical treatments mainly relieve symptoms of raised ICP from impaired CSF drainage but do not address the underlying cause. Future research should be based on the latest understanding of the pathophysiology of IIH, with the aim of further identifying the primary causes and developing more refined, targeted treatments rather than merely providing symptomatic treatment.
A landmark study by Iliff et al.2) used a murine model and in vivo 2-photon microscopy to measure vessel wall movement within the PVSs during each cardiac cycle, thus quantifying vessel wall pulsatility. They visualized perivascular CSF influx into brain tissue and established arterial pulsatility as a key driver of CSF-ISF exchange. By manipulating cerebrovascular pulsatility—either reducing it through internal carotid artery ligation or partially occluding the brachiocephalic artery, or augmenting it pharmacologically—they demonstrated altered glymphatic CSF tracer influx, linking reduced arterial pulsatility to CSF transport into the brain.2)
As discussed in the “Anatomy and Physiology” section, glymphatic clearance relies on arterial and venous pulsatility, intact PVSs, arteriovenous pressure gradients, and AQP4 channels—factors that can be disrupted by systemic, inflammatory, or focal diseases. In steno-occlusive cerebrovascular disease, severe arterial stenosis may dampen penetrating artery pulsatility, thus potentially impairing glymphatic clearance. This mechanism could explain cognitive decline in patients with significant supraaortic stenoses and its improvement following revascularization.
Cardiogenic dementia51) is defined as cognitive deterioration due to heart diseases such as heart failure, myocardial infarction, and atrial fibrillation, and is believed to be present in 20% to 80% of all patients with acute or chronic heart failure52) with milder forms of cognitive impairment being more often seen compared to more severe forms. In addition to the above-mentioned diseases, coronary artery disease is also associated with the development of cognitive impairment53) due to reduction of myocardial contractility. Various factors have been implicated in the development of cardiogenic dementia, including systemic low-grade inflammation, systemic oxidative stress, and BBB injury.51) In addition, it is hypothesized that the reduction of cerebral blood flow leads to cognitive deterioration; however, the underlying mechanism for the development of cardiogenic dementia has yet to be fully elucidated. It has been proposed that chronic cerebral hypoperfusion induces the development of neurodegeneration, in particular since chronic cerebral hypoperfusion was correlated with microinfarcts and the development of CAA and Aβ aggregates.54) With increased understanding of the GS, this phenomenon may be explained as Alzheimer type neuropathology that can be promoted through cardiogenic hypoperfusion by activating the amyloidogenic pathway and inhibiting Aβ clearance. In addition to decreased pulsatility due to heart failure, reduced myocardial contractility, or atrial fibrillation, chronic cerebral hypoperfusion induces retention of CO2 leading to cerebral vasodilation, which in turn will decrease the size of the PVSs and decrease pulsatility, all of which lead to decreased CSF egression into the interstitium.55)
The cognitive decline observed in patients with severe symptomatic internal carotid artery (sICA) stenosis also was previously poorly understood. While previous studies linked it to impaired cerebral blood flow (CBF) affecting metabolism, the precise changes in CBF, white matter, and metabolism before and after carotid endarterectomy (CEA) remained unclear.56) Animal models of cerebral hypoperfusion, such as carotid artery ligation in rodents, have demonstrated accelerated amyloid plaque deposition, CAA progression, and cognitive decline in Alzheimer’s models.2) Increasing evidence suggests Alzheimer’s disease and cerebrovascular cognitive impairment share underlying mechanisms, including hypoperfusion, BBB dysfunction, inflammation, and hemorrhagic susceptibility.12)
The role of arterial pulsatility in CSF-ISF exchange suggests that age-related or disease-related declines in pulsatility may contribute to inefficient interstitial waste clearance, including amyloid, further linking cerebrovascular dysfunction to neurodegeneration.57)
In addition to medium and large vessel atherosclerotic steno-occlusive disease, cerebral small vessel disease (CSVD) leading to leukoencephalopathy will also have a profound effect on the GS and has recently been implicated in cognitive deterioration.58) Traditionally, it has been thought that the brain damage present in CSVD was a consequence of cerebral hypoperfusion and hypoxia; however more recent evidence suggests that CSVD leads via impairment of the GS, to cognitive decline. Arteriolosclerosis, caused by hypertension or diabetes mellitus, is the underlying pathology in CSVD and leads to “stiffening” of the arteries, which will cause decreased pulsatility in the brain, thus likely also impairing glymphatic flow.58) In addition to sharing the same underlying mechanism, CSVD can actually further impair glymphatic function due to its histological substrate, i.e., microinfarctions.59) These microinfarctions can create focal areas of glymphatic dysfunction due to CSF tracers being trapped within areas of microinfarction, suggesting further lack of convective flow and waste clearance.59)
The GS’s convective clearance mechanism relies not only on arterial pulsatility, which facilitates CSF movement from PVSs into the interstitium, but also on venous pressure. Elevated venous pressure reduces the arteriovenous gradient, thereby impairing convective flow through neural tissue. Since most CSF efflux occurs along intramedullary or transmedullary veins, arterialization of these veins is likely to have more severe consequences than arterialization further downstream in the venous system or superficial venous channels.2)
AVMsIn unruptured brain AVMs, case reports have documented hydrocephalus due to hydrovenous disorders or hydrodynamic disequilibrium, leading to reflux into transmedullary veins.60) This hydrodynamic concept of hydrocephalus suggests that, in addition to absorption through arachnoid granulations (Pacchionian bodies), CSF is also reabsorbed via physiological transependymal flow into the transmedullary veins. Increased venous pressure in these veins, caused by arteriovenous shunting, may hinder CSF absorption by diminishing the pressure gradient between the capillary bed and surrounding CSF spaces.
In 1 patient reported in the literature, cognitive decline and symptoms resembling normal pressure hydrocephalus were reversed following partial embolization, which reduced transmedullary arterialization. This led to significant improvements both clinically and on imaging.60)
In neonates and infants, particularly those with vein of Galen AVMs or high-flow pial shunts, this mechanism is well documented. At this developmental stage, arachnoid granulations are immature, making CSF absorption entirely dependent on hydrodynamic mechanisms. This explains why in this age group there is nearly always a hydrodynamic disequilibrium present, which leads to hydrocephalus that is not responsive to external ventricular drainage but only to decreasing the venous pressure by treating the vascular malformation.61)
More recently, a retrospective study demonstrated dilated PVSs to be present surrounding brain AVMs, particularly within the perinidal region of unruptured and untreated brain AVMs.62) These dilated PVS were characterized by T2 hyperintensity with CSF suppression on FLAIR sequences. The authors noticed that they corresponded in location to the intraoperative observation of a “shiny plane” between the AVM nidus and brain parenchyma. The PVSs were predominantly present around the nidus, but not surrounding the proximal feeding arteries or distal draining veins. One may presume that the increased size of the PVSs surrounding the AVM nidus is related to a local imbalance of the GS due to locally increased venous pressure leading to “backflow” into the perinidal arterial PVSs, thus leading to dilatation of these structures.62)
dAVFsdAVFs, another form of brain arteriovenous shunting, have also been associated with cognitive decline. While previously attributed to venous congestion and ischemia in deep brain structures such as the bilateral thalami,63) recent evidence suggests impaired CSF reabsorption and glymphatic dysfunction may also play a role. A case series by Brito et al. reported that all 6 patients with cognitive decline in their study exhibited either reflux and arterialization of transmedullary veins or venous congestion.63) Cognitive function improved after shunt obliteration.
Furthermore, a prospective study assessing cognitive function before and after dAVF treatment found that even patients without overt cognitive symptoms showed cognitive improvement posttreatment.64) This suggests that mild venous hypertension in transmedullary veins can subtly affect cognition. These findings align with the Neuropsychology in dural Arterial Fistula study, a multicenter prospective cohort investigation, which found significant cognitive improvements in attention, executive function, memory, and language following endovascular embolization of dAVFs.65) Another study by the CONDOR Consortium focused on patients with cognitive impairment, though it lacked a standardized neuropsychological assessment, potentially limiting comparability.66)
By including all patients, regardless of initial cognitive symptoms, these studies highlight the prevalence of subclinical cognitive deficits in dAVF patients and emphasize the importance of routine cognitive assessments. The observed cognitive improvements following normalization of the arteriovenous pressure gradient suggest that enhanced waste clearance may be the cause of this amelioration, further linking the GS, cognitive performance, and neurovascular diseases.
The glymphatic network serves as a unique and vital system for circulating ISF throughout the central nervous system, making its evaluation a promising frontier in the neurosciences. In the neurovascular field, including its recent expansions in treating intracranial hypertension, we believe that there are various points of interaction with the GS. In this review, we described that blockage of the paraarterial spaces due to SAH, abnormalities in AQP concentrations as a potential underlying cause of IIH, changes in pulsatility due to steno-occlusive diseases, and decreased arteriovenous pressure gradients due to shunting diseases all interfere with the GS at various levels. Understanding the GS may therefore not only deepen our understanding of neurovascular diseases but also help us in the future to identify new treatment avenues.
We wish to acknowledge Dan Andreae and Patricia Holt-Hornsby for their support.
The authors declare that they have no conflicts of interest.
