Skip main navigation

Not Just Blood: Brain Fluid Systems and Their Relevance to Cerebrovascular Diseases

Originally published 2022;53:1399–1401

Stroke research has largely focused on blood supply to the brain, particularly via large arteries leading from the heart, the muscular intracranial large arteries, large cerebral veins and venous sinuses, and more recently perforating arterioles, capillaries, and venules. However, these blood vessels only serve part of the brain’s fluid management, nutrient delivery, and waste clearance system. The other, and until recently largely neglected, aspect of brain fluid and waste management is the system that flushes the brain, draining interstitial and cerebrospinal fluid (CSF) and waste from the cranial cavity. Details of this non–blood-vessel brain circulation are incomplete, but there is now enough clinical relevance for it to be the focus of this Advances in Stroke: Diagnosis and Imaging, particularly as key elements are now visible on routine magnetic resonance imaging (MRI).

Brain Fluid Pathways: a Short Summary

Brain interstitial fluid is critical for maintaining normal neuronal and glial function. Some interstitial content forms from selective fluid and molecule transfer at the blood-brain barrier (BBB) and cell metabolic activity, but most is now thought to come from influx of CSF via perivascular spaces (PVS).1

CSF is initially formed by the choroid plexus in the ventricles, exits the fourth ventricular foramina, and bathes the brain surface; thence, it is thought to enter PVS around perforating arterioles. At the capillary neurogliovascular unit level, it is thought that astrocyte aquaporin 4 (AQP4) molecules help transfer such fluid into the interstitium.1 Fluid plus waste then exit the interstitium along perivenular spaces continuing as perivenous spaces around cortical veins over the brain surface, which remain invested in a layer of meninges, thus separating dirty interstitial effluent from clean CSF (Figure [A and B]). This peri-arteriole-to-interstitial-to-perivenous fluid system is called the glymphatic system.


Figure. Perivenous, mlymphatic, and cerebrospinal fluid (CSF) waste drainage routes. T1-weighted (A), fluid attenuated inversion recovery (FLAIR) axial magnetic resonance imaging (MRI) near vertex (B), and FLAIR of sagittal sinus (C) near the torcula 20 min after intravenous gadolinium shows gadolinium-enhanced blood in cortical veins and the sagittal sinus (A; white arrows), in brain interstitial fluid draining in perivenous spaces (B; ochre arrows; white arrows=sagittal sinus) and (C) in mlymphatics (C; long yellow arrows) running along the sagittal sinus (C; white arrow). D, Heavily T2-weighted coronal MRI of sagittal sinus near vertex without intravenous gadolinium shows mlymphatics (long yellow arrows) alongside the sagittal sinus (white arrow). E, Axial FLAIR MRI (without intravenous or intrathecal contrast injection), inferior frontal fossa, showing increased CSF signal in inferior frontal sulci immediately superior to the cribriform plate: even very small amounts of protein or cellular debris increase FLAIR CSF signal at this key cranial fluid drainage point.5

Historically, CSF was thought mainly to drain via dural arachnoid granulations to dural venous blood, but it now seems that most of the perivenular dirty effluent enters meningeal lymphatics (mlymphatics), which travel along the major venous sinuses (Figure [C and D]) to exit the skull. CSF may also enter directly into other mlymphatics around cranial nerves, principal arteries, and skull base dura including cribriform plate2 (where nasal lymphatics intertwine closely with olfactory nerves), which exit the skull, continuing as lymphatics in the neck to cervical lymph nodes.

In addition to microvascular level AQP4 suction, the movement of fluid into PVS, and of CSF from the ventricles to the brain surface, is thought to depend on vessel pulsation during the cardiac cycle, vasomotion, CSF pressure gradients, and respiratory effort. With each heartbeat, CSF oscillates between the intracranial and perispinal subarachnoid spaces3 and is encouraged along the periarteriolar PVS into the brain1; vasomotion also pushes fluid along the PVS4; respiratory motion (eg, deep breath) may add by transiently increasing cranial venous efflux and CSF influx.

Interestingly, periarteriolar PVS,2 cortical perivenous spaces, mlymphatics, and waste CSF draining at the cribriform plate5 are all visible on routine MRI (Figure). Perivenous spaces and mlymphatics are better seen after intravenous gadolinium, since the gadolinium passes via CSF and interstitium to the perivenous and mlymphatic channels but can be seen on noncontrast fluid attenuated inversion recovery (FLAIR; waste including proteins elevates fluid signal)6 and some T2-weighted sequences. PVS count, volume, and other characteristics (shape and location) can be quantified,7 making it now possible to assess and understand brain fluid management in health and disease.

Relevance to Cerebrovascular Disease

A few PVS are normal on MRI but increase in number with age, stroke risk factors, particularly hypertension,2 small vessel disease features, particularly white matter hyperintensities, advanced cerebral amyloid angiopathy, and genetic small vessel diseases like cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, and in neurodegenerative disorders including Alzheimer disease (AD).

Risk factors may act via changing vascular stiffness. Blood pressure elevation changes the normal smooth, continuous forward motion of fluid along the PVS to jerky and intermittent, with reduced net forward flow, facilitating debris deposition, obstructing the PVS.1 Increased systemic and intracranial pulse pressure, for example, with hypertension, are associated with increased PVS visibility and reduced CSF motion.2 Some white matter hyperintensities appear to form around PVS,8 suggesting impaired tissue clearance may contribute to white matter hyperintensities. In cerebral amyloid angiopathy, PVSs increase in number and colocate with amyloid deposition,9 possibly reflecting failed protein clearance. Increasing age may link PVS to proteinopathies (eg, cerebral amyloid angiopathy and AD) through falling CSF production leading to reduced brain flushing.1 Inferior frontal sulci CSF signal (debris) increases on FLAIR with age and PVS (Figure [E]).5 Glymphatic system activity may increase during sleep, promoting waste clearance in rodents,1 perhaps explaining the importance of sleep for brain health; in humans, there is limited evidence of PVS associating with poor sleep habits, while pathological sleep (eg, obstructive sleep apnea) increases stroke and small vessel disease.10 Thus, increased numbers and size of PVS on MRI indicate impaired fluid and waste clearance and worsening brain health.

In acute ischemic stroke, rodent experiments suggest that CSF is quickly taken up into PVS in the affected artery territory, and may contribute to, and offer alternative targets to reduce, infarct edema.11 However, this does not explain several well-established observations: the earliest edema is not interstitial but cellular due to cell membrane ion pump failure; consistent with this, the earliest MRI sign of ischemia is restricted diffusion due to cell swelling restricting interstitial fluid movement. Alternatively, vasospasm occurring shortly after arterial occlusion could increase CSF uptake into the PVS while the arteriole is in spasm. So, while CSF may be a source of fluid entering the interstitial space and cells in acute stroke, it may be an epiphenomenon rather than primary cause of swelling.

Relevance in Chronic Disorders

Normal pressure hydrocephalus, where there is ventricular enlargement, white matter hyperintensities, and cognitive and gait impairments, is notoriously difficult to manage, and precise causes remain elusive. In normal pressure hydrocephalus, delayed brain MRI after injection of gadolinium into spinal CSF shows slow uptake of gadolinium into PVS and delayed brain clearance.12 Recently, refractory hydrocephalus after subarachnoid hemorrhage was successfully treated by inserting a tiny CSF-to-dural-venous valve shunt13 (an arachnoid granulation mimic) via retrograde venous catheterization, to release CSF into venous blood above prespecified pressures. Minimally invasive intracranial vascular access is now routine in stroke through daily use for thrombectomy and aneurysm coiling: perhaps several disorders of CSF regulation like normal pressure hydrocephalus might in future be managed by minimally invasive insertion of tiny CSF-to-dural-venous shunts.


Although visible on routine daily MRI, a major part of the brain’s fluid and waste management system, closely interwoven with and interdependent on blood vessels, has been misunderstood or ignored. In addition to small vessel disease, in which sporadic, hemorrhagic, and genetic forms show signs of PVS (glymphatic) dysfunction, impaired waste drainage occurs in several common neurodegenerative proteinopathies, and impaired CSF management in other chronic disabling disorders. PVSs should be seen as early markers of adverse brain health, including of sensitivity to or poor control of risk factors. Research should address controversies over interstitial fluid sources, drainage routes, methods to quantify FLAIR signal (excess debris), and mlymphatics. Perhaps arachnoid granulations act as safety valves (suggested by successful insertion of small CSF-dural-venous valves to treat hydrocephalus13) rather than main drainage routes. Deeper understanding of human glymphatic and mlymphatic function during sleep, of ways to enhance function, and of impacts of routine stroke preventions on these systems, could offer enhanced ways to protect brain health, providing resilience to stroke and dementia.

Article Information

Disclosures Dr Liebeskind is a consultant as imaging core laboratory for Cerenovus, Genentech, Medtronic, Rapid Medical, and Stryker. The other author reports no conflicts.


The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

For Sources of Funding and Disclosures, see page 1401.

Correspondence to: Joanna M. Wardlaw, MD, Centre for Clinical Brain Sciences, 49 Little France Crescent, Edinburgh EH16 4SB, United Kingdom. Email


  • 1. Rasmussen MK, Mestre H, Nedergaard M. Fluid transport in the brain.Physiol Rev. 2021. doi: 10.1152/physrev.00031.2020CrossrefGoogle Scholar
  • 2. Wardlaw J, Benveniste H, Nedergaard M, Zlokovic B, Mestre H, Lee H, Doubal FN, Brown R, Ramirez J, MacIntosh BJ, et al. Perivascular spaces in the brain: anatomy, physiology, and contributions to pathology of brain diseases.Nat Rev Neurol. 2020; 16:137–153. doi: 10.1038/s41582-41020-40312-zCrossrefGoogle Scholar
  • 3. Wardlaw JM, Benveniste H, Williams A. Cerebral vascular dysfunctions detected in human small vessel disease and implications for preclinical studies.Annu Rev Physiol. 2022; 84:409–434. doi: 10.1146/annurev-physiol-060821-014521CrossrefGoogle Scholar
  • 4. van Veluw SJ, Hou SS, Calvo-Rodriguez M, Arbel-Ornath M, Snyder AC, Frosch MP, Greenberg SM, Bacskai BJ. Vasomotion as a driving force for paravascular clearance in the awake mouse brain.Neuron. 2020; 105:549–561.e5. doi: 10.1016/j.neuron.2019.10.033CrossrefMedlineGoogle Scholar
  • 5. Zhang JF, Lim HF, Chappell FM, Clancy U, Wiseman S, Valdés-Hernández MC, Garcia DJ, Bastin ME, Doubal FN, Hewins W, et al. Relationship between inferior frontal sulcal hyperintensities on brain MRI, ageing and cerebral small vessel disease.Neurobiol Aging. 2021; 106:130–138. doi: 10.1016/j.neurobiolaging.2021.06.013CrossrefGoogle Scholar
  • 6. Albayram MS, Smith G, Tufan F, Tuna IS, Bostanciklioğlu M, Zile M, Albayram O. Non-invasive MR imaging of human brain lymphatic networks with connections to cervical lymph nodes.Nat Commun. 2022; 13:203. doi: 10.1038/s41467-021-27887-0CrossrefGoogle Scholar
  • 7. Ballerini L, Booth T, Valdés Hernández MDC, Wiseman S, Lovreglio R, Muñoz Maniega S, Morris Z, Pattie A, Corley J, Gow A, et al. Computational quantification of brain perivascular space morphologies: associations with vascular risk factors and white matter hyperintensities. A study in the Lothian Birth Cohort 1936.Neuroimage Clin. 2020; 25:102120. doi: 10.1016/j.nicl.2019.102120CrossrefGoogle Scholar
  • 8. Huang P, Zhang R, Jiaerken Y, Wang S, Yu W, Hong H, Lian C, Li K, Zeng Q, Luo X, et al. Deep white matter hyperintensity is associated with the dilation of perivascular space.J Cereb Blood Flow Metab. 2021; 41:2370–2380. doi: 10.1177/0271678X211002279CrossrefGoogle Scholar
  • 9. Charidimou A, Farid K, Tsai HH, Tsai LK, Yen RF, Baron JC. Amyloid-PET burden and regional distribution in cerebral amyloid angiopathy: a systematic review and meta-analysis of biomarker performance.J Neurol Neurosurg Psychiatry. 2018; 89:410–417. doi: 10.1136/jnnp-2017-316851CrossrefGoogle Scholar
  • 10. Huang Y, Yang C, Yuan R, Liu M, Hao Z. Association of obstructive sleep apnea and cerebral small vessel disease: a systematic review and meta-analysis.Sleep. 2020; 43:zsz264. doi: 10.1093/sleep/zsz264CrossrefGoogle Scholar
  • 11. Mestre H, Du T, Sweeney AM, Liu G, Samson AJ, Peng W, Mortensen KN, Stæger FF, Bork PAR, Bashford L, et al. Cerebrospinal fluid influx drives acute ischemic tissue swelling.Science. 2020; 367:eaax7171. doi: 10.1126/science.aax7171CrossrefMedlineGoogle Scholar
  • 12. Eide PK, Ringstad G. Delayed clearance of cerebrospinal fluid tracer from entorhinal cortex in idiopathic normal pressure hydrocephalus: a glymphatic magnetic resonance imaging study.J Cereb Blood Flow Metab. 2019; 39:1355–1368. doi: 10.1177/0271678X18760974CrossrefGoogle Scholar
  • 13. Lylyk P, Lylyk I, Bleise C, Scrivano E, Lylyk PN, Beneduce B, Heilman CB, Malek AM. First-in-human endovascular treatment of hydrocephalus with a miniature biomimetic transdural shunt [published online December 3, 2021].J NeuroInterv Surg. 2021. doi: 10.1136/neurintsurg-2021-018136CrossrefGoogle Scholar


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.