Cerebrospinal Fluid Homeostasis and Hydrodynamics: A Review of Facts and Theories

Cerebrospinal fluid (CSF) presents a clear, colorless liquid, that fills the neuroaxis’ cavities and surrounds its’ tissue. This essential fluid participates in a significant number of homeostatic processes, including nutrients transportation, the removal of catabolic waste products, weight reduction through buoyancy (brain weight <50 g), hydromechanical stress absorption, thermal stress regulation, and neurotransmitter transportation [1]. It consists predominantly (99%) out of water, while the remaining 1% accounts for electrolytes, proteins, neurotransmitters, and glucose [2]. Cells are expected to be absent, although a minor population consisted out of white blood cells (<5/mL) may be introduced indirectly [3]. The volume of the CSF is expected to be constant and estimated at 150 mL (125 mL located in the subarachnoid space [SAS] and 25 mL in the ventricular system). This volume is strictly defined by equal production and reabsorption rates of approximately 18–25 mL/h or 430–530 mL/day in adults [4], ensuring the regulation of the intracranial pressure (ICP), known to be dependent upon its combination with the intracranial blood and brain tissue volumes (Cushing formula for the Monroe-Kellie hypothesis [5]). It is assumed that CSF is mainly produced by the choroid plexuses (CPs), before circulating through the brains’ ventricular system, its’ cisterns, and the SAS, to be finally reabsorbed, predominantly into the venous sinuses at the level of arachnoid granulations (AGs). This doctrine, which was initially described by von Luschka [6] and Faivre [7] and was later supported by experimental data, mainly by the observations of Dandy on obstructive hydrocephalus [8] and the consequences of choroid-plexectomy in canines [9], is still quoted, and is even employed as a foundation upon which the definition of hydrocephalus was based [10], influencing and defining treatment techniques and strategies, without being subjected to any major corrections since its’ introduction. However, this seems to be a rough simplification of a very complex system, considered by some researchers to be just an unproven assumption [11], founded upon experimental models that could not be reproduced by a significant number of researchers [12-14], due to the biased methodology initially employed [15]. Taking into consideration the increasing number of studies defying the exact sites and physiological mechanisms of CSF production, absorption, and circulation, this review attempts to provide a holistic picture of the proposed regulatory apparatus theories, pointing toward the key developments supporting or defying them.

Major databases were included in the conduction of the study, among them MEDLINE, CENTRAL, and EMBASE. Gray literature was screened employing a manual hand-search technique. No restrictions were imposed regarding the time or type of publication, including all studies posing at least an abstract in English. The following search terms were used as a free text and MeSH terms: “cerebrospinal fluid,” “CSF,” “hydrocephalus,” “choroid plexus,” “choroid villi,” “arachnoid granulations,” “arachnoid villi,” “meningeal lymphatic vessels,” “aquaporins,” “choroid plexectomy,” “choroid ablation,” “choroidectomy,” “choroid plexus cauterization,” “endoscopic third ventriclostomy,” “ETV,” “CPC,” “Bulat Klarica Oreskovic,” “Weed Dandy Cushing.”

CNS and its’ interstitial fluid are effectively protected by the blood-brain-barrier (BBB), a single semipermeable layer of microvascular endothelial cells, fused by tighter junctions than that expressed in other membranes, thus restricting potential free access of systematic fluid [16]. The process of CSF secretion has been extensively investigated. The majority of authors advocate the classical Weed-Dandy-Cushing hypothesis; however, others embrace an alternative view.

According to the classical theory, CSF is actively formed, predominantly by the CPs,floating inside the lateral, third, and fourth brain ventricles, while other production sites, including the parenchyma and the ependyma, are believed to contribute at a significantly lesser extent [17]. These formations are highly vascular, containing a significant number of blood-vessel projections (villi), embedded within their stroma. Each of these projections consists out of a large central capillary with a fenestrated leaky endothelium, covered by a cuboidal epithelium, arranged in a single layer, and secured by tight junction proteins, forming the blood-CSF barrier (Fig. 1). This structure exhibits functions that are restricted by the BBB, including the entrance of selective inflammatory cell populations to the CNS [18], and the formation of CSF. The last is supposed to be a two-phase process. Initially, the osmotic gradient difference leads to the passive filtration of plasma across the fenestrated capillaries’ endothelium and into the interstitial space, producing a plasma ultrafiltrate, which is then actively transported into the ventricular space, throughout the epithelium [17, 19-21]. The epithelial tight junctions of the CPs among others, contain a significant proportion of claudin-5 transmembrane proteins, known to be permeable for water molecules [22, 23]. It should be highlighted that brain capillaries, unlike the common and CPs capillary network, possesses a tight endothelium, lacking any permeable transmembrane proteins, restricting any potential fluid exchange between the compartments [24], thus their role in the creation of CSF was rejected from an early stage (Fig. 2).

The expectation that a miniature structure (2–3 g [25]) could be responsible for the homeostasis of CSF producing approximately 500 mL/day, at a rate higher than many secretory epithelia (0.2 mL/min^–1 per g) [4] may be considered absurd. However, the presence of highly vascular tissue projections ensures the existence of a significant contact surface area, calculated at approximately 220 cm³. The regional blood flow is believed to be 10 times greater than that found in other regions, including the cortex, expressing a 3–5 times faster movement [26]. The circulating blood has a 1.15 times higher hematocrit [27] than the one included in the systematic arterial blood. Mathematical models taking into consideration the aforementioned data combined with the estimated regional average arterial flow (2.86 μL/mg/min) discovered an estimated CSF secretion rate that is roughly equal to the expected absorption rate [28], something that was utilized as a definitive evidence of the classical theory. Experimental models employing isolated extra corporally perfused CPs provided support to the aforementioned data [29, 30]. Other similar studies performed either on animal or on human specimens reported an almost identical chemical composition between the CSF and the secretions of the isolated CPs, while the Nernst electrochemical ion gradient equation analysis based on Na⁺, K⁺, Cl⁻, and divalent ion concentrations provided further support to the CPs origin of the CSF [31-33].

According to the aforementioned, it is expected that choroid plexectomy would significantly reduce the formation of CSF. However, the employment of this method in hydrocephalus patients, presented disappointing results, with literature suggesting an overall production rate reduction of up to 40% in experimental animal models [34, 35], thus it was abandoned, at least as a single approach technique. Some authors point toward a potential beneficial effect when combined with endoscopic third ventriculostomy [36]; however, the most up to date systematic review with meta-analysis failed to provide supportive evidence [37], while the latest study of the Hydrocephalus Research Network goes one step further suggesting a potentially negative effect to the overall outcome in pediatric patients [38].

The secretion of CSF is presented as an energy-consuming process, thus it is expected to be pressure independent [39-41], at least till the 30 cm H₂O threshold, believed to be capable of dysregulating active processes [42]. Various animal experimental models presented a potentially notable reduction of CSF production rates at pressures remarkably below the stated values, defying the significant pump-like action theory [43, 44], also associated with the circulation of CSF. The fact that the contact surface of the cerebral capillaries is approximately 5,000 times more extensive than that of the CPs [45, 46] has driven some authors to propose that the formation of CSF takes place throughout the entire cerebral capillaries network instead of the CPs alone.

Developmental observations in humans and other mammals have highlighted that the secretion of CSF is present even before the formation of the CPs anlagen [47, 48]. Anatomical studies in fishes of the elasmobranch superorder, mainly sharks failed to detect a visible communication between the intraventricular CSF and the extraventricular one, while a number of lower vertebrates lacking CPs were found to possess ventricular cavities filled with CSF [19, 47, 49].

Experimental data on felines presented that there is no leak when an artificial working channel is introduced in the cerebral aqueduct, after its acute blockage, a fact that was interpreted as proof of absence of an active process during secretion, bringing into question even the circulation of CSF through the ventricular system [50], at least in the unidirectional way in which it is presented. Moreover, further investigation in similar models, when the aqueduct of Sylvius was obstructed did not result in ventricular dilatation, leading authors to defy the classical hypothesis [51].

CPs may be considered as the main site of production, but they are certainly not the only one. The volume produced by them is assumed to reach approximately 60%–85% of the total CSF production, while it has been confirmed that a significant portion is formed outside the CPs.

Ependyma. It is believed that the extra-choroidal portion mainly originates from the interstitial fluid produced by ependymal cells [4, 29, 31, 52], at a rate up to 30% [53], with even higher rates presented in a felines’ spinal cord ependyma model [54]. The majority of these experiments, however, were criticized [19], due to the drastic procedures employed, that could potentially influence the results, leading to biased conclusions.

The Importance of Water-Channels. The role of water channels, known as aquaporins (AQP), in the passive water movement is well documented. Experimental results have failed to achieve a consensus regarding the identification of the exact isoform contributing to the formation of CSF. Some authors have identified the importance of AQP-4 [55], an isoform that is profusely found in the subpial and perivascular spaces of astrocyte terminals, the external and internal limiting membranes of glial cells and the basolateral membrane of ependymal cells [56-59]. Others have highlighted the role of AQP-1 [60], which is expressed by the CPs [61, 62] and the leptomeningeal vasculature cells [63], while the equal contribution of both forms cannot be excluded [64, 65]. Other endothelial water-transport proteins, expressed either on the luminal, the apical, or on both of the membranes, including predominantly GLUT1, NKCC1, and Na⁺/K⁺/2Cl⁻ cotransporter, are known to contribute at a significant extent [66]. These channels ensure the transport of water alongside other substances, even against osmotic gradients, providing a significant volume of bidirectional dynamic water movement, despite their minor individual net flow. It must be highlighted that the subject of mediated brain water movements, under physiological conditions, is a matter of ongoing research.

The contradicting data presented afore, led some researchers to defy the conventional theory and to propose alternative homeostatic mechanisms of CSF regulation. After more than 3 decades of conducting animal experimental models and reviews, a group of authors concluded that the volume of CSF is dependent upon hydrostatic and osmotic gradients and that a rapid and constant fluid exchange takes place between CSF, ISF, and plasma throughout the entire CNS capillaries network [67]. This is based upon the assumption that the SAS is not isolated from the brain’s parenchyma by any type of barrier, and that the Virchow-Robin paravascular space is able of ensuring a free fluid exchange between the CSF and the parenchyma, something that has been previously defied. According to these novel understandings hydrocephalus, thus elevated ICP and dilatation of the ventricles, is the result of a disruption in osmolarity and blood vessel permeability, leading to a compensatory water movement. These can be associated with various pathophysiological states, including inflammatory processes, intraventricular bleeding, presence of neoplasms, malfunction of water channels, and damage to the BBB [11]. If these statements are valid, then they could potentially explain hydrocephalus cases where no obvious obstruction can be identified [51, 68-71]; however, this would mean that the actual hydrocephalus definition should be significantly modified, while treatment strategies should probably be reconsidered.

Classical hypothesis advocates the unidirectional cranio-caudate circulation of CSF, from the lateral ventricles [52], where active secretion (volume) alongside the CPs vascular structures pulsations manifest a pump-like activity, forcing the liquid through the ventricular system, an action which is accommodated at a significant extent by the synchronous beating of the ependyma’s cilia [72]. The cilia’s movements are repeated at a speed of 28–40.7 Hz and are regulated by serotonin, ATP, the pituitary adenylate cyclase-activating polypeptide, the melanin-concentrating hormone, and dopamine [73-78]. After passing through the aqueduct of Sylvius and the fourth ventricle, CSF is forced through the lateral foramen of Lushka into the SAS of the cisterns and over the cortex, or through the medial foramen of Magendie to the spinal SAS. As presented this model resembles at a significant extent the blood and lymph circulations, thus it is occasionally referred to as the “third circulation.”

Recent experimental models defy this view [79]. A systematic review of human studies attempting to define the normal intracranial and intraspinal pressures in human subjects, registered an ICP of 5.9–8.3 mm Hg in the upright and 0.9–16.3 mm Hg in the supine position, while the intradural pressure of the lumbar region was found to be 7.2–16.8 mm Hg in the upright and 5.7–15.5 mm Hg in the supine position. It would be paradoxal for a fluid to flow from a site without any gradient (0 cm H₂O in the cranial region) to a region showing higher pressure (+30 cm H₂O in the lumbar region) as presented in other models too [80]. According to the aforementioned, the unidirectional circulation is not possible, at least in the upright position. It is more likely that CSF is slowly moving bidirectionally in accordance to the systolic-diastolic displacement causing dispersion of macromolecules. Some authors highlight the importance of differentiating the movement of CSF from the distribution/redistribution of substances, pointing toward a misinterpretation of experimental results due to confusion between these two processes, when the last is employed as a sign of circulation.

The absence of a unidirectional flow brings into question the widely accepted assumption of hydrocephalus being associated with an interruption in communication between the site of secretion and that of the absorption, on which the definition of this entity is based [81]. Literature highlights that there is a possibility of aqueductal obstruction occurring as a repercussion of hydrocephalus instead as a causative factor. A rodent model of mice infected by reovirus-1 provided support to the aforementioned, as it was observed that progressive dilatation of the ventricular system, and the resulted midbrain compression led to the stenosis of the aqueduct [82]. This is further supported by a couple of additional rodent models [83, 84]. Literature also includes a case of an MRI-confirmed near-obstruction aqueductal stenosis in a female lacking any imaging or clinical signs of hydrocephalus, during a 5-year follow-up [71].

Cine-phase-contrast MRI studies have been regularly employed as a definitive proof of flow through the Sylvian aqueduct. This technique is known to provide quantitative blood flow information [85]. However, the average flow in healthy adults was found to be significantly higher than the proposed rates of CPs production [86]. Additionally, a retrograde caudo-cranial flow was registered in children <2 years old [86] and in patients with normal pressure hydrocephalus [87, 88]. The validity of the technique is questionable, as its algorithm is founded upon the hypothesis of the unidirectional cranio-caudate circulation, and the assumption that the diameter of the passage remains constant. Literature further highlights that this evaluation is subjective and significantly dependent upon the technical limitations of the imaging system and the acquisition parameters employed [89].

Recent experimental models additionally defy the presence of the unidirectional flow [2, 90-93]. If this is true, the CSF does not circulate and substances migrate across anatomical regions due to the stirring action associated with the well-known pulsatile movements.

It is generally accepted that after flowing through the SAS, CSF is passively reabsorbed through arachnoid membrane projections, known as Pacchionian-arachnoid villi (AV) and into the dural sinuses (mainly the superior sagittal sinus) and the lateral lacuna [31, 94]. These anatomical structures (AV) are composed out of a dense collagen connective tissue and clusters of arachnoid cells. They presumably act as a unidirectional pressure-dependent valve (Fig. 3), with an opening pressure of 3–5 mm Hg [95], a statement also supported by electron microscopy studies noting the presence of pressure-sensitive pores [96, 97]. CPs are considered to be an accessory reabsorption site contributing in the reuptake of up to 10% of the overall production, presenting a function resembling that of a proximal renal tube [19, 98-100].

The exact mechanism of absorption is not yet fully understood. Anatomical studies on isolated dural tissue segments acquired from primates and canines have pointed toward the presence of open channels within AV, suggesting the existence of a free bulk flow [101, 102]. However, electron microscopy investigation brought into question the complete penetration of the dural sinuses by AV, highlighting that a tight junctions barrier could be present between the endothelial cells restricting any potential free passage [103-105]. The non-fenestrated layer was even registered under elevated ICP, despite the distension of the anatomical structures. The aforementioned led the authors to conclude that an active process was required in order to transport high molecular weight substances, thus micropinocytosis was proposed. This is however unlikely, as physiological studies presented a fast and unrestricted movement of macromolecules, while the transport was present even post mortem, facts arguing against an active process [94, 106, 107]. A recent review highlights that literature has yet to demonstrate anatomical structures actually extending into the dural sinuses, something that would come in contrast with the physiological observations of the unrestricted outflow [108]. Although physiological evidence supports the valve function theory, one of the earliest controversies was the lack of AV in humans and some other species (rat and sheep) before parturition, a state that could even persist for a significant period of time after birth [109-112]. Moreover, a recent study, employing a 3T MRI technique has shown an age-related variation in the number, size, and distribution of AV in the superior sagittal sinus and bony structures, presenting that a significant portion of people of various age groups lack completely or express a minimum number of granulations, with individuals >60 years presenting them predominantly in the cranial osseous matter [113]. If these structures are so important for the reabsorption process, why are these subjects not presenting with signs of hydrocephalus? It is only logical to conclude that AV do not contribute significantly to the absorption of CSF and to classify them as accessory pathways (Fig. 4).

The confirmation of meningeal lymphatic vessels being present in the human and primate dura [114-117] has recently rekindled interest in the potential importance of this system in the reabsorption of CSF, predominantly the part located near the sagittal and transverse sinuses, and alongside the skull base vascular network. The exact potential mechanism of this homeostatic process remains to be confirmed. The arising question regarding the potential participation of this system in the overall absorption process comes from the presence of an arachnoid barrier layer between the SAS and the dura [118, 119], known to be formed out of multiple layers of epithelial cells connected with tight junctions [120], thus restricting any potential water movement. A number of experimental rodent models [121, 122] provide further evidence against this probable reabsorption site, leading to the belief that it presents an accessory pathway at the best [123]. However, some authors driven by strong evidence pointing toward the presence of a significant (up to 48%) reabsorption through the cranial lymphatic system and further flow through the deep neck nodes, highlight the possibility of this being the primary structure of CSF drainage [112, 124-126]. Literature proposes various potential routes of communication between the SAS and lymphatic system, with the one alongside cranial nerves, being the most likely.

Animal models have traced a movement along the olfactory bulb, passing through the cribriform plate into the nasal mucosa and through the nasal lymphatics, alongside pharynx’s vessels to the deep cervical lymph nodes [123, 127-130]. This was also confirmed in humans [126, 131-133]. The importance of this pathway was demonstrated by obstructing it in ovis neonates, an act that resulted in an evident disruption of the CSF absorption [134]. Despite the understanding of the importance of this system, the exact anatomical route by which CSF reaches lymphatic vessels is still questionable, with literature proposing three probable mechanisms. The first assumes that fluid and solutes may access the interstitial space, thus lymphatic vessels of the submucosa passing by loosely adhered perineural cells, in a complete absence of any form of a barrier [135]. The second theory implies that perineural cells may exhibit some barrier properties similar to the meninges [136]; however, this would require an active transfer process, thus it is unlikely for the same reasons stated afore regarding the AV. The third assumption is founded upon rodent model results suggesting that lymph vessels may project through the cribriform plate accessing the SAS directly [114, 130, 137-139], providing immediate access to the CSF. Opponents of these theories base their belief upon potential faulty methodology employed during the conduction of the experimental procedures mainly associated with the pressure applied during the introduction of tracers that could result in the potential rupture of fine barrier layers.

Experimental models have proposed the potential movement of tracer substances between the outer fibrous and the inner pia matter layer covering the nerve itself [121, 128-130, 140-143]. However, the histological and ultrastructural examination of the orbit was not able to confirm the route connecting the optic nerve sheath to the lymphatics, with conjunctival ones being proposed as the most likely [144].

Additional routes alongside other cranial nerves may theoretically present similar functions. Several animal studies have implied a potential implication of the trigeminal (V) [118, 121, 142], facial (VII) [121, 122], and vestibular (VIII) [145, 146] nerves; however, there was no strong evidence to support these statements. According to literature the implication of the jugular foramina is more likely, advocating its’ association with the glossopharyngeal (IX), vagus (X), and accessory (XI) nerves [121, 142], a fact that remains to be further investigated.

Research regarding CSF absorption in the spine has been scarce. The leading opinion is that this anatomical regions’ contribution is limited and certainly inferior to that of the cranial region [147-149]. Potential loci include the so-called subarachnoid angle, located at the region of spinal nerve root formation, especially their dorsal part. Literature suggests that these extensions of the SAS contribute to the drainage either through their connection to AV [150], and associated venous plexuses, or to lymphatics [151].

The current knowledge of CSF dynamics led to the proposal of several theories regarding the pathophysiology of hydrocephalus. According to the bulk flow theory, introduced by Walter Dandy, hydrocephalus is the result of a blockage in the circulation of CSF. If the obstruction is located prior to the level of the AGs, the term obstructive hydrocephalus applies, whereas if there is no evident barrier, the disruption is attributed to an insufficiency of the absorptive mechanisms, a state known as communicating hydrocephalus. This highly accepted classification, however, is not taking into consideration the impact of arterial pulsations. The CNS should not be simplistically regarded as a laboratory tube containing a mixture of solids and fluids passing through. The interactions between the brain and cord parenchyma, the CSF and the blood that enters the cranial cavity via a vigorously pulsatile flow are extremely complicated. According to our knowledge, Edgar Bering was the first scientist taking into account these pulsations [152]. He believed that the CPs are the main conveyors of the arterial pulses into the cerebrum. This normal occurrence has a pivotal role in the dilatation of the ventricular system. In cases of abnormal ICP or decreased tissue tolerance, the impact of the transmitted waves on the ventricle walls may present a hummer-like effect, contributing to their enlargement. He tried to prove this by performing a unilateral plexectomy in an animal model after inducing hydrocephalus. This resulted in the enlargement of the contralateral ventricle, which also presented with a higher pressure amplitude. Two decades later, Di Rocco et al. [153] showed that the introduction of a pulsating balloon into an ovine ventricle will result in its dilatation, providing further support to this theory. Greitz [154] conducted a cine-MRI study, presenting that alterations in brain compliance, attributed to various pathological states (e.g., post hemorrhagic or post infectious), may potentially lower the ability of the parenchyma to accommodate the entering arterial pulsations, causing a decompensation of the parenchyma/fluid system, introducing the terms restricted arterial pulsation hydrocephalus or increased capillary pulsation hydrocephalus. Preuss et al. [155] formulated the pulsatile vector theory, describing the “shockwaves” produced during each heart circle, including the arterial pulsations, the brain pulsations, and their reflections (e.g., from the ventricular walls or the cranial vault). According to their theory, a possible asynchronous or misbalanced relationship between these waves can lead to ventricular distention.

As we learn more about hydrocephalus, the bulk flow model appears oversimplistic but, interestingly, we still use the concept of obstructive and communicating hydrocephalus during the surgical management of these patients. For example, an adult with a tectal-plate tumor and triventricular hydrocephalus is an excellent candidate for ETV, as the resulting flow diversion is expected to bypass the mechanical obstruction, thus resolve the elevated ICP. The employment of ETV success score (ETVSS), introduced by Kulkarni et al. [156], also provides support to this theory, as the probability of success is calculated to be 80%. On the other hand, an adult with postinfectious, thus communicating hydrocephalus, is not expected to substantially benefit from a CSF flow diversion, as the obstruction is located at the level of AV. The ETVSS confirms once again this theory, with success rates being calculated at 50% therefore, the insertion of a VP shunt as an initial treatment is justifiable.

However, according to the novel understanding, obstruction is not the only cause of insufficient absorption. Another factor significantly influencing the overall outcome is the age. A newborn (<30 days) presenting with aqueductal stenosis and treated with ETV has half the chances of success, compared to an adult with similar pathology, according to the ETVSS. The literature also includes a multicenter study regarding the management of infants with aqueductal stenosis, presenting significantly better results in those older than 6 months when compared to a younger population of similar characteristics [157]. Employment of the bulk flow model attributes these observations to the immaturity of the absorptive mechanisms, including the absence of AV, alongside the insufficient resorption through the venous part of the parenchymal capillaries and the dural lymphatic system. The inadequate arterial and parenchymal pulsations absorption should also be considered. This phenomenon is associated with an inadequate compliance of the poorly myelinated white matter of the brain in these age groups, and the buffering action caused by the open sutures and fontanelles, especially the anterior one, due to its close relationship with the third ventricle, which diminishes the craniocaudal pulsation waves. These waves are known to originate from the vascular structures located inside the brain parenchyma and ventricles, before being propagated throughout the intracranial tissue and fluid to the calvarium, where it is amplified and reflected downward, forcing the CSF to move through the foramina of Monro and toward the floor of the third ventricle, increasing the chances of ETV success [155].

A similar example where a dual explanation may apply is the increased success rates of ETV after cauterization of the CP in hydrocephalic children. Warf [158] showed that combining ETV with cauterization of the CP of both lateral ventricles had better results than ETV alone. A simplistic explanation according to the bulk flow model is that the elimination of the major CSF production source results in a downregulation of its overall production volume, something expected to subvene the immature absorption system in achieving satisfactory results. The removal of CPs is also expected to significantly decrease the pressure amplitude of the intraventricular pulsations, further preventing ventricular dilatation, as shown by the previously cited Bering experiment. But there is a possibility that an exactly opposite mechanism can apply. Their overall configuration (resembling that of a sea sponge or soft coral) may act as a pillow and pressure buffer that absorbs the traveling pulsatile vectors into the CSF-filled ventricles. The CSF is regarded incompressible, as a fluid, and therefore transmits the pressure waves directly and unchanged on the ventricular walls. By removing the CPs (cauterization) these pulse waves, remain unbuffered and contribute more efficiently to the CSF movements throughout the ETV stoma.

It becomes evident that despite the updated theories and the increasing knowledge regarding the field of CSF circulation and hydrocephalus, a single convincing answer or a unified theory is still absent. The pathophysiology of hydrocephalus is multifactorial and both bulk flow and the novel theories have strong and weak points. A combination of their more convincing aspects should apply in order to illuminate the numerous dark angles of hydrocephalus development and management.

Neither of the proposed assumptions may be accepted or rejected with certainty. The CSF dynamics are definitely based on more complex processes than previously understood. CPs contribution to the secretion and absorption is evident. However, there is a continuous fluid exchange taking place between CSF, ISF, and plasma throughout the entire CNS capillaries network. The unidirectional circulation may be absent, while the solute migration throughout the entire CNS compartments system may be a product of stirring action associated with the fro and to movements. The role and importance of AG should be further evaluated. Hydrocephalus is the result of disruption between the secretion and absorption volumes, potentially associated with hydrostatic and osmotic disbalances. The role of the pulsatile waves originating from the intracranial vascular structures should not be neglected. The importance of obstruction is questionable, as it may present the result of the elevated ICP, being associated with the pressure applied to the anatomical structures. If this is the case the entire hydrocephalus classification should be reconsidered.

This review is part of the doctoral dissertation entitled “Microanatomical, imaging, and neuroendoscopic correlations of the 3rd ventricle floor,” conducted by the leading author at the Aristotle University of Thessaloniki, School of Medicine. The supervisor of the thesis is the senior author Prof. Christos Tsonidis, who, alongside Prof. Konstantinos Natsis and Ass. Prof. Konstantinos Kouskouras, forms the advisory committee.

Statement of Ethics is not applicable.

The authors have no conflict of interest to declare.

The study, similarly to the doctoral dissertation including it, is completely funded by the leading author.

Marios Theologou conceived the original idea, conducted a literature search, and wrote the draft of the manuscript, under the guidance and help of Prof. Konstantinos Natsis, Ass. Prof. Konstantinos Kouskouras, and Assoc. Prof. Fotios Chatzinikolaou. Panagiotis Varoutis, Nikolaos Skoulios, and Vassilios Tsitouras conducted proofreading and assisted to a substantial extent in the creation and editing of the included figures. Prof. Christos Tsonidis provided guidance and encouragement while supervising the work and conducting proofreading. The manuscript was discussed prior to submission between the authors. All authors contributed to the final manuscript.