Pseudotumor Cerebri Pathophysiology

Authors

  • Brian E. McGeeney MD, MPH,

    Corresponding author
    1. Neurology, Boston University School of Medicine, Boston, MA, USA
    • Address all correspondence to B.E. McGeeney, Neurology, Boston University School of Medicine, 715 Albany Street, Boston, MA 02118, USA.

    Search for more papers by this author
  • Deborah I. Friedman MD, MPH, FAAN

    1. Neurology & Neurotherapeutics and Ophthalmology, University of Texas Southwestern Medical Center, Dallas, TX, USA
    Search for more papers by this author

  • Conflict of Interest: None.

Abstract

Pseudotumor cerebri syndrome (PTCS) is an uncommon disorder of raised intracranial pressure of unknown etiology. The signs and symptoms have been well described but the pathogenesis remains a mystery. Most of the evidence suggests increased resistance to cerebrospinal fluid outflow as being pivotal to the disorder. Any comprehensive theory on causation will have to explain the preponderance of obese women of childbearing age with primary PTCS and lack of ventriculomegaly in the disorder. It is possible that female sex hormones, along with endocrinologically active adipose tissue, directly result in the syndrome, in those genetically predisposed. Aldosterone has been proposed also as important in the development of PTCS. Vitamin A, in the form of retinoic acid, may also play a pivotal role, and is influenced by both estrogen and adipose tissue. This article reviews proposed mechanisms of PTCS.

Abbreviations
ANP

atrial natriuretic peptide

AQP

aquaporin

AVP

arginine vasopressin peptide

BNP

B-type natriuretic peptide

CNP

C-type natriuretic peptide

CP

choroid plexus

CSF

cerebrospinal fluid

ICP

intracranial pressure

IIH

idiopathic intracranial hypertension

MRA

magnetic resonance imaging

PA

per annum

PTCS

pseudotumor cerebri syndrome

RBP

retinol binding protein

Pseudotumor cerebri is a syndrome of raised intracranial pressure (ICP) without ventriculomegaly, tumor or mass intracranially, producing signs and symptoms of raised intracranial pressure such as headache, photophobia, nausea, pulsatile tinnitus, transient visual obscurations, visual field defects, papilledema, and the feared visual loss. Cerebrospinal fluid (CSF) composition is normal. Sometimes known as idiopathic intracranial hypertension (IIH), the pseudotumor cerebri syndrome (PTCS) nomenclature is more inclusive of secondary etiologies. IIH, which is the primary, or idiopathic form of PTCS generally affects obese females of childbearing age and has an estimated incidence of 1.6/100,000 per annum (PA), but rises to 11.9/100,000 PA in obese women.[1] The signs and symptoms of PTCS are well characterized, but our understanding of pathophysiology is lacking, despite many putative associations. Any credible theory of causation will have to explain the preponderance of obese women with IIH. Men can develop IIH, albeit rarely, and it behooves the treating clinician to look exhaustively for secondary causes in males. Cerebrospinal fluid dynamics and homeostasis in PTCS are complex and incompletely understood. Treatments are aimed at reducing cerebrospinal fluid production, reducing CSF pressure on the optic nerves and lowering CSF pressure via a drainage procedure. There are no good clinical studies providing evidence for the standard treatments. The Idiopathic Intracranial Hypertension Treatment Trial is the first large-scale, randomized, double-masked study to determine whether acetazolamide is superior to placebo in subjects with mild visual loss who are treated with a supervised weight reduction diet.[2] The primary outcome for the study is perimetric mean deviation, with results to be released in 2014. The secondary objective of the trial will study genetic factors related to vitamin A and hormones implicated in the etiology IIH.

Criteria for the diagnosis of PTCS require the presence of papilledema, and a CSF pressure over 250 mm CSF[3] in adults and over 280 mm CSF in children.[4] Pressures less than 200 mm CSF are considered normal, and those between 200 mm and the appropriate cut-off value for age are regarded as non-diagnostic. The CSF contents must be normal, and neuroimaging shows no evidence of a tumor, infection, or inflammation. Subtle signs of intracranial hypertension, such as an empty sella, cerebellar tonsillar descent, flattening of the posterior sclerae, tortuosity and distention of the subarachnoid perioptic space, protrusion of the optic nerve papillae into the vitreous, and transverse venous sinus stenosis are frequently seen.[5] The diagnosis may be made with relative certainty when these features are present, particularly in the typical patient demographic. Bouchat is credited with the first description of a case of pseudotumor cerebri syndrome in 1866, expressed in Passot's thesis of 1913, as reported by Johnston and colleagues.[6] Passot hypothesized hyper-production of CSF as causative in this disorder.[7] Hughlings Jackson provided one of the earliest descriptions in English when describing a young woman with recoverable optic neuritis from 1880.[8] Pseudotumor cerebri was characterized more fully by Quincke, who developed the lumbar puncture needle pivotal to the diagnosis, and by Nonne who used the term “pseudotumor cerebri” in his 1914 paper.[9, 10] Symonds proposed “either an excessive secretion from the choroid plexus or a defective absorption through the subarachnoid space” to explain PTCS in a 1931 paper on “otic hydrocephalus.”[11] Normal ventriculography in PTCS was first described by Davidoff and Dyke in 1937 and separately by Dandy also in 1937, when the treatment of choice was subtemporal decompression.[12, 13] Dandy was the first to postulate the role of the venous system in PTCS to explain the lack of ventriculomegaly.[13]

By the 1950s the clinical features were well defined, and multiple etiological factors were being proposed.[14] There is now a long list of possible risk factors with limited and varying degrees of evidence, including head injury, hematological disorders like anemia, polycythemia and pro-thrombotic disorders, endocrinological disorders such as polycystic ovarian syndrome, the plethora of cerebral venous outflow obstruction causes, medications like tetracyclines, lithium, synthetic growth hormone, thyroid hormone replacement in children, vitamin A derivatives, and family history among others. All of these alleged associations are thought to disrupt CSF dynamics but only uncommonly, making it difficult to demonstrate a relationship from standard epidemiological studies with moderate numbers of subjects. This manuscript will focus on the idiopathic variety of PTCS.

Cerebrospinal Fluid Lifecycle

About two thirds of the CSF is produced by the choroid plexus (CP) epithelium in the cerebral ventricles, and most of the remainder comes from the ependymal lining of the ventricles.[15] There is evidence that pressure sensitive elements in the CSF axis produce peptides (such as atrial natriuretic peptide) that can regulate the production of CSF, but this process has not been properly elucidated.[16] Approximately 600 mL of CSF is produced daily, but synthesis is notably reduced in later life,[17] possibly contributing to why IIH is a disorder of younger people only. The production of CSF involves energy dependent Na+/K+-ATPase situated on the luminal membrane of the CP epithelium transporting Na+ from the epithelial cytoplasm into the CSF space. Other ion movements (such as K+, Cl, HCO3) follow the electrochemical gradient into the CSF via transporters. Large amounts of water cross the epithelium, facilitated by aquaporin-1 channels. The carbonic anhydrase family of enzymes produces protons and bicarbonate which are used by the transporters at both the CSF facing and basal epithelial membrane. Acetazolamide, by inhibiting carbonic anhydrase, reduces available ions, raises cytosolic pH and slows passage of Na+ into the CSF.[18]

Cerebrospinal fluid flows out of the ventricular system and into the subarachnoid space where it is resorbed. It is classically taught that CSF is resorbed solely through herniations of arachnoid membrane into the venous sinuses called arachnoid granulations (with microscopic villi), but there are other important pathways for CSF outflow. CSF crosses the arachnoid epithelium in intracellular vacuoles rather than between cells.[19] Flow of CSF across the arachnoid villi is proportional to the pressure difference between CSF space and venous sinuses, and inversely proportional to resistance to flow across the arachnoid villi. However, other mechanisms are likely involved in CSF resorption besides the arachnoid granulations. Two studies failed to show arachnoid granulations in autopsy specimens of fetuses, but clearly before birth there is a CSF circulation, and it is likely that arachnoid granulations only appear around the time of birth. Investigations have demonstrated drainage of CSF into the head and neck lymphatic system, despite a lack of intracranial lymphatic vessels.[20] A major site of CSF drainage is through the cribriform plate, traversed by olfactory nerves each surrounded by a sheath CSF and dura.[21] The CSF drains into nasal lymphatics and in larger mammals such as sheep, the capacity of nasal lymphatics to drain CSF is about equal to the arachnoid granulations and is presumed to occur in humans also. Radiolabeling techniques have demonstrated that significant spinal absorption occurs in humans, which can amount to 50% of total CSF resorption.[22] The various CSF outflow pathways make increased production of CSF in PTCS less plausible as there are plenty of exit routes.

Aquaporins

Aquaporins (AQP) are water permeable channels (mostly excluding ions) that in many cells provide the main route for bidirectional water movement across the membrane, by facilitating the osmotic movement of water. At least 9 different aquaporins have been identified in the central nervous system (CNS).[23] Aquaporins are not expressed in excitable cells but are found in supporting cells. AQP1 is found in the choroid plexus epithelium and AQP4 is the preeminent water channel in the human brain, mainly expressed in glial cells. Cerebrospinal fluid production is reduced by 25% in AQP1 knockout mice.[23] Brain AQP4 is strongly expressed at the borders between brain parenchyma and major fluid compartments – astrocyte foot processes (blood–brain–barrier), glia limitans (brain-subarachnoid CSF), and ependymal cells (brain-ventricular CSF). Water flows between compartments in response to osmotic and hydrostatic forces. AQP4 facilitates water flow into and out of the brain. As aquaporins have a major role in water movement between the major compartments, anything that alters their function may play a role in PTCS. Interestingly, acetazolamide has been shown to inhibit AQP4 activity[24] and may also modulate AQP1.[25] Topiramate and zonisamide, antiepileptic medications which inhibit carbonic anhydrase to a lesser extent than acetazolamide, also have an inhibitory effect on AQP4 that approaches that of acetazolamide.[26] However, a genetic association study that sequenced the gene for AQP4 on chromosome 18 in 28 patients with IIH and 52 control subjects found no association with AQP4 gene variants and IIH.[27] APQs are nonetheless attractive candidates in the pathogenesis of PTCS and require further study.

Cerebrospinal Fluid Dynamic Changes in Pseudotumor Cerebri

As intracranial pressure is increased, it may be concluded that the rise in pressure is originating in at least one of the compartments – CSF, blood, interstitial fluid, or brain cells. CSF production is relatively constant and independent of CSF pressure, at least up to a point, hence increases in pressure are not quickly countered with reduced production of CSF, making the development of PTCS easier once the process begins.[28-30] An increased production of CSF has been suggested in PTCS,[31] but there is no supportive evidence for this or any structural changes in the choroid plexus, such as hypertrophy. There is some evidence to suggest that CSF production is normal in PTCS.[32] Most of the focus in PTCS has been on the resistance to CSF absorption. There have been a number of studies involving CSF infusion, and all demonstrated increased resistance to CSF absorption in PTCS.[33-35] Gjerris found abnormally low CSF conductance in 12 of 14 patients with PTCS, with the 2 normal values measured after starting treatment.[34] Janny demonstrated resistance to flow in all 16 patients studied with PTCS.[35] While most studies have focused on the intracranial CSF compartment, recent magnetic resonance imaging (MRI) evidence using dynamic phase contrast techniques suggests that there is increased extraventricular CSF volume and decreased jugular venous outflow in IIH compared to control subjects.[36] Altered spinal canal compliance may be a contributing factor.[37] It seems likely that there is impaired absorption in PTCS as core to the pathophysiology, possibly due to a field effect involving epithelial membranes which would involve all the routes of CSF outflow.

The Venous Sinuses

It has long been recognized that major flow obstructions in the cerebral venous system, like transverse sinus thrombosis related to ear infection could cause PTCS.[14] An important question is whether there are primary abnormalities of the venous system in PTCS.[38] King first proposed venous stenosis as etiologic in PTCS but retracted his claim in a follow-up publication when he demonstrated abolition of a venous pressure gradient in the transverse sinus after CSF drainage, suggesting that parts of the venous system are collapsible under high pressure, and resolve with pressure normalization.[39, 40] Not everyone agreed with King's updated view and continued to propose that venous narrowing was a primary problem in PTCS.[41] Filling defects may be difficult to discern when using magnetic resonance angiography (MRA) 2D time-of-flight method but are well visualized on 3D contrast-enhanced MRA images, the latter imaging techniques indicating a high prevalence of abnormalities with PTCS. Farb found substantial bilateral venous sinus stenosis in 27 of 29 patients with PTCS using modern MR imaging techniques.[42] Owler and colleagues summarized studies on venous sinus compression in humans with PTCS, noting venous sinus compression is a possible result of increased intracranial pressure but not invariably so.[38]

Anatomical asymmetries of the venous sinuses are important, with the right transverse sinus typically dominant and containing more intraluminal septae which can cause filling defects. In one study, 59% of the subjects had an aplastic or hypoplastic left transverse sinus,[43] a congenital variant that is sometimes mistakenly interpreted as acquired stenosis. Large arachnoid granulations may also cause filling defects which are preferentially found in the distal and middle thirds of the transverse sinus.[44] Such granulations have been implicated in causing PTCS, especially when involving a dominant transverse sinus, the other being hypoplastic.[45] A study of dogs with occlusion of one or both transverse sinuses concluded that obstruction of one transverse sinus had minimal effect on CSF pressure while occlusion of both transverse sinuses produced elevated CSF pressures.[46]

Attention has focused on bilateral transverse sinus stenosis rather than unilateral stenosis as being hemodynamically significant. Whether venous abnormalities are a primary or secondary phenomenon or not, the flow restriction and resulting back pressure may further increase CSF pressure and perpetuate the situation.[47] Rohr, in a volumetric MRI study, demonstrated compression of the entire dural sinus tree in those with PTCS.[48] Sudden reduction of CSF pressure with lumbar puncture likely reverses flattened venous sinuses and improves the CSF outflow dynamics resulting in a prolonged benefit, even though CSF is replaced quickly. This is also an argument in favor of the controversial practice of venous sinus stenting in PTCS.[49]

Similarly, the degree of recanalization and collateral circulation in the presence of a stenotic or obstructed dominant venous sinus was investigated in a detailed review of MR venograms in pediatric patients being evaluated for suspected IIH.[50] The authors reviewed 145 cases, of whom 27 had contemporaneous CSF opening pressures and compared them to 50 age-matched control subjects in whom IIH was not suspected. The diagnosis was based on ICD codes, which are not highly reliable for this purpose,[51] and the charts were not reviewed to determine whether or not the patients truly had IIH. About half of cases showed single or multiple sites of venous outflow obstruction in the dominant-side circulation in suspected IIH cases, most commonly in the transverse sinus but also in the sigmoid sinuses and jugular bulbs. Non-physiological collateral circulation was seen in 68% of cases with dominant-side venous outflow obstruction. Collateral recanalization was present in the 17 of 20 cases, having a CSF opening pressure greater than 20 cm CSF (mean pressure was 31 cm CSF). Only 2 of 50 control subjects had obstruction of the dominant circulation, one of whom had evidence of collateral circulation. Unfortunately, the authors provided no data on whether or not the patients had papilledema, or whether they had an acute or chronic course. The development of collateral vessels in response to venous sinus obstruction implies some degree of chronicity. Bono and colleagues reported bilateral transverse sinus stenosis in patients with migraine and tension-type headaches who do not have papilledema.[52, 53] Subjects with bilateral stenosis were much more likely to have raised CSF pressure compared to controls, despite normal optic discs and were diagnosed with “IIH without papilledema.” These studies need to be replicated by other researchers, and the full implications of these findings remain unclear at present. It is uncertain whether or not these patients fall into the PTCS spectrum or have a different condition. The entity of PTCS without papilledema does exist, but is uncommon and may be over-diagnosed in the anxious patient with an erroneously high-CSF pressure reading from Valsalva maneuver during the procedure.[54, 55] There may be other confounding factors, such as use of opioids, performance of the LP under sedation, and possibly an influence of chronic head pain, leading to increased CSF pressure readings in these cases.[56] In summary, the venous sinuses are undoubtedly affected in PTCS, but generally as a secondary phenomenon, and the venous sinus theory does not readily encompass the female predominance of PTCS in adults and adolescents.

It has also been theorized that patients with IIH have a propensity to thrombophilia, perhaps increasing the risk of venous sinus thrombosis or occlusion. One study found an increased prevalence of the C677T methylenetetrahydrofolate reductase (MTHFR) mutation (38% of patients vs 14% of controls), and postulated that thrombophilia-hypofibrinolysis and subsequent thrombosis, particularly when associated with obesity, may be promoters of IIH.[57] Obesity is associated with hypofibrinolysis through increased plasminogen activator inhibitor activity, resulting in a prothrombotic state.[58] One study found increased fibrinogen levels and red blood cell aggregation in patient with IIH compared to control subjects.[59] There are no pathological studies demonstrating thrombotic occlusion, however, except in cases of thrombosed sinuses.

Sodium and Water Regulation: Aldosterone, Vasopressin and Natriuretic Peptides

Aldosterone is a mineralocorticoid with well-known actions on the kidneys (eg, regulation of sodium homeostasis), but it is also active on epithelial cells of the choroid plexus, serving to enhance the activity of the Na+/K+-ATPase exchanger on the luminal membrane.[60] Enhanced Na+ passage into the CSF results in increased CSF production. Two cases of primary aldosteronism associated with PTCS were reported in 2002, and there have since been other reports of PTCS with primary and secondary hyperaldosteronism.[61] Interestingly, the initial description of primary hyperaldosteronism at the Central Society for Clinical Research meeting in 1954 occurred in “middle-aged” (30-50 years) women with a long history of headache, and a subsequent literature review of 145 cases found that 75% were women, 54% of whom had headaches that were not further described.[60, 62] As yet there is no direct evidence of increased CSF production in cases of hyperaldosteronism, although aldosterone is found in the cerebrospinal fluid in levels correlating to plasma aldosterone and has a known effect on the regulation of CSF volume.[60, 63] However, polycystic ovarian syndrome (PCOS) is associated with hyperaldosteronism. One study comparing recumbent and upright levels of aldosterone between subjects with PCOS and controls matched for age and body mass index (BMI) found a highly significant increase in PCOS subjects' levels (P < .001) with values 1.5-2 times higher than in controls.[64] This finding was corroborated in a second study by different investigators.[65] PCOS has been linked to IIH anecdotally, as well as in a study of 65 women with IIH and 102 female control subjects.[57] Fifty-seven percent of the patients with IIH had PCOS as defined by standard criteria. Almost all of the 37 patients with IIH and PCOS were either obese (BMI 30-39 kg/m2, 43%) or extremely obese (BMI >40 kg/m2, 51%), in contrast to the 28 women with IIH who did not have PCOS (57% obese, 21% extremely obese).

Orthostatic (idiopathic) edema is characterized by dependent edema that resolves during recumbency in the absence of cardiac or renal disease, and abnormal retention of sodium, water or both. The condition occurs almost exclusively in women and was demonstrated in 80% of 29 female patients with IIH.[66] Hyperaldosteronism in the upright posture was found in patients with idiopathic edema.[67] Abnormalities in aldosterone homeostasis would potentially explain both the female sex predilection and relationship to obesity found in patients with IIH, has potential therapeutic implications, and warrants further exploration.

Arginine vasopressin (AVP) is a peptide hormone that regulates body water in the nephron and is also found in the CNS. Elevated CSF AVP levels were demonstrated in IIH patients, approaching values twice that of controls, which did not correlate with lumbar puncture opening pressure.[68] There was no difference in CSF oxytocin levels or plasma levels of either peptide hormone. Intracerebroventricular infusion of AVP increased ICP in conscious goats.[69] Other studies have confirmed increased CSF AVP levels in IIH but also demonstrated similar findings in other disorders associated with increased ICP. Thus, AVP is likely released as a response to increased ICP and is not specific to IIH.[70]

The natriuretic peptides antagonize the renin-angiotension-aldersterone system. Atrial natriuretic peptide (ANP) and B-type natriuretic peptide (BNP) plasma concentrations are inversely proportional to BMI.[71] BNP and C-type natriuretic peptide (CNP) receptors are found in the CP and intraventricular administration of ANP reduces elevated ICP. A study of IIH patients found no difference in precursor peptides to ANP, BNP, or CNP although plasma levels of BNP and CNP were lower in IIH subjects than in controls, not correlating with BMI or CSF pressure.[71]

Obesity

The relationship of obesity to idiopathic PTCS is well known, but the mechanism of causality is uncertain. Obesity is rampant in many industrialized countries and, although the incidence of IIH is increasing, it is still a rare disorder. Thus, it seems unlikely that obesity alone is causative. It was postulated that there is increased intra-abdominal pressure in obese individuals, transmitted via venous vessels to the spine and resulting in increased intracranial pressure.[72] This theory is not borne out by the science. In the presence of obesity, intra-abdominal pressure is increased in the prone position against a rigid surface, but two recent studies comparing CSF pressure measurement in the prone vs the lateral decubitus position found little difference related to body position.[73, 74] Moreover, pregnancy, in which the gravid uterus compresses the inferior vena cava, is not associated with an increased risk of IIH.[75] Even if it were true, it does not explain the marked gender difference in PTCS.

A Scottish study looking at 254 patients (disorders associated with increased CSF pressure were excluded) noted normal individuals with pressures up to 280 mm CSF.[76] The correlation of CSF pressure with weight was very small (correlation coefficient 0.19). Curiously, Bono did not identify any obese people with a CSF pressure above 200 mm CSF in a study involving 51 overweight and obese subjects, all having a normal cerebral magnetic resonance venography (MRV).[77] Obesity under normal circumstances contributes very little if anything to elevated CSF pressure. However, obesity is a common factor in women with PTCS who have no other identified etiological risk factor. In 3 case-control studies, only obesity and recent weight gain were identified as associated with PTCS.[78-80] Recurrence of symptoms in IIH is associated with weight gain.[81] The IIH Treatment Trial may provide further insight, with prospective data regarding the effect of dietary intervention with and without treatment with acetazolamide in patients with mild visual loss.[2]

Adipose tissue is an active endocrine organ secreting many different substances, from pro-inflammatory cytokines to adiponectin, and the link between obesity and PTCS is likely contained somewhere among these agents. It is less likely that PTCS and obesity are independent manifestations of a common endocrinological cause. It seems unlikely that raised ICP causes obesity but it has been proposed.[82] Obesity is also a risk factor for progression from episodic to chronic migraine for uncertain reasons.[83] Adipose tissue contains aromatase, which may be a link between obesity and PTCS. Aromatase produces estrogens from plasma androstenedione and is more prevalent in the buttock regions (female distribution) than in abdominal fat.[84] The activity of aromatase in adipose tissue clearly increases with advancing age.[85] This age-related increase does not fit with the incidence of PTCS, but the dramatic fall of female sex hormones at menopause may negate the increased aromatase activity. Also, adipocytes secrete potent mineralocorticoid-releasing factors[86] resulting in increased aldosterone levels. The excess mineralocorticoid activity as noted earlier may be a risk factor for PTCS. A fourth possible link between obesity and PTCS involves retinol, which is converted to retinaldehyde and retinoic acid in adipose tissue. Retinoic acid is associated with PTCS as will be discussed below.

Obstructive sleep apnea, often affiliated with obesity, has been independently associated with PTCS, especially in men.[87, 88] It has been postulated that the hypoxia and hypercapnia result in cerebral vasodilation causing an increase in ICP, which is sustained if sufficient venous sinus compression ensues. One recent small retrospective case-controlled study failed to find obstructive sleep apnea as an inherent risk factor for PTCS.[89]

Vitamin A

One of the most interesting and well-established aspects of PTCS is the association with vitamin A. Vitamin A regulates growth and differentiation in most cells of the body and its physiological concentration in plasma is regulated within a narrow range.[90] Retinol is bound specifically to retinol binding protein (RBP), which is primarily synthesized in the liver but also in adipose tissue. After cellular uptake, most of the retinol returns to the plasma, and a small proportion is converted into retinoic acid, its active metabolite.[91] The mechanism of cellular uptake is not well understood but appears to be at least partially mediated by a retinol binding protein receptor. Radiolabeled iodinated RBP and RBP receptors bind with high affinity in the choroid plexus in human brain and animal models.[90, 91]

There is good evidence linking hypervitaminosis A and PTCS, and one can reliably induce the syndrome if enough vitamin A is ingested.[92, 93] Gerrit de Veer first described the severe illness with headache accompanying the toxic ingestion of vitamin A rich polar bear liver in 1596 while on a voyage to find the Northeast Passage to Asia.[94] At the time, the cause of the acute illness was unknown. Vitamin A derivatives such as isotretinoin, etretinate, and the leukemia drug all-trans-retinoic acid all can cause PTCS.[95] The relationship of vitamin A to PTCS is complicated, however, as both hyper- and hypovitaminosis A are associated with PTCS, although the majority of the literature concerns hypervitaminosis A. There is a strong association with hypovitaminosis A and raised CSF pressure in pigs, dogs, calves, and lambs.[96] Panozzo described a 27-year-old woman who, 5 years after gastric bypass surgery for obesity, developed clinical hypovitaminosis A including PTCS which disappeared after restoration of normal vitamin A blood levels.[97]

Our understanding of the mechanism of action of vitamin A-induced PTCS was advanced by Calhoun who, using a ventriculo-cisternal perfusion technique, demonstrated that the cause of increased CSF pressure in hypovitaminosis A is impaired CSF absorption, similar to the infusion studies on PTCS.[98] Additionally, there is some evidence that the impaired CSF absorption was associated with structural changes in and adjacent to arachnoid villi. In vitamin A deficient calves, Hayes found fibrosis of the interstitium of arachnoid villi with increased collagen providing a pathological basis for the raised ICP.[99] Adipose tissue not only stores vitamin A, but is the site where retinal is converted to its active metabolite retinoic acid. Among other actions, retinoic acid induces the synthesis of neurosteroids such as progesterone in the CNS, activating the mineralocorticoid receptor.[100, 101]

Jacobson and colleagues found significantly elevated serum retinol levels in a total of 16 women with PTCS, compared to 70 healthy controls.[102] Warner studied vitamin A levels in CSF in three groups of subjects, those with PTCS, those with raised ICP from other causes, and those with normal ICP.[103] There was a significantly higher concentration of vitamin A in the CSF of those with PTCS only. A follow-up study compared the retinol/retinol binding protein (RBP) ratio in those with and without PTCS.[104] The retinol/RBP ratio was greater in the CSF than in the serum, especially in subjects with PTCS, suggesting more free retinol may be toxic and may interfere with CSF homeostasis. Retinoic acid exerts its effects by controlling the expression of over 500 genes.[105] Intriguingly, retinoic acid has been shown to affect endothelial cell function and influences the composition and functional properties of their underlying extracellular matrix. These actions could theoretically link excess retinoic acid to PTCS but does not explain the female predominance in IIH.[106] Estrogen has also been shown to induce retinoic acid synthesis. Another risk factor for PTCS that might act by altering endothelial permeability is tetracycline use. Doxycycline has been shown to prevent vascular permeability factor/vascular endothelial growth factor induced vascular permeability.[107] Overall, there are multiple lines of evidence linking retinol with PTCS which will be investigated further as part of the IIH Treatment Trial.[2]

Corticosteroids

The effect and role of glucocorticoids and mineralocorticoids is complex. The association between Addison's disease and raised ICP was reported in 1952, followed by reports of PTCS after exogenous corticosteroid administration.[108, 109] Cushing's disease is also associated with PTCS.[110] Johnston and colleagues identified 1013 cases of PTCS with a likely known etiology, and steroids were implicated in 6.1% (70 patients).[111] Corticosteroids (essentially glucocorticoids) may precipitate PTCS during prolonged courses and, perhaps more frequently, during withdrawal.[112] However, corticosteroids as a treatment for PTCS have been used since the 1960s. It has been clearly demonstrated that administration of corticosteroids is associated with a considerable reduction in CSF production.[113, 114] Importantly, cessation of corticosteroid therapy is associated with a marked increase in CSF resistance to absorption,[115] which explains its association with PTCS. The reduction in CSF production with corticosteroids is likely due to reduced activity of Na+/K+-ATPase,[116] but this does not explain how prolonged treatment can be causative, and for that we have to look for other effects of steroids. Corticosteroids also activate mineralocorticoid receptors on the choroid plexus, stimulating the production of Na+/K+-ATPase pumps, and in this way could induce PTCS. This seems unlikely, as it invokes increased production of CSF, and idiopathic PTCS is mostly associated with reduced CSF outflow. There are likely other important effects of corticosteroids on CSF dynamics that are not known. Although corticosteroids have a therapeutic role for acute PTCS emergencies, their side effects do not warrant long-term therapeutic use in PTCS.

Sex Hormones

As IIH predominantly affects post-pubertal females, one must strongly suspect female sex hormones as being etiologic. The absolute effect of female sex hormones as a risk for PTCS in most people is likely small given the uncommon frequency of IIH overall, and that may be why three studies failed to find a link between estrogen-containing contraceptives and PTCS.[78-80] Much larger sample sizes would be needed to show an effect. The widespread use of estrogen-containing oral contraceptives worldwide and relative rarity of PTCS suggests that these medications are not the sole provoking factor. The risk that estrogen poses for PTCS goes beyond the prothrombotic risk of cerebral venous thrombosis. As yet, there is no good experimental or epidemiological study demonstrating the link between estrogen and PTCS. Fluctuations in intracranial pressure have been suggested to occur during the menstrual cycle, pregnancy, and during the use of oral contraceptives. However, although PTCS occasionally begins or recurs during pregnancy, there is no increased risk of PTCS occurring during pregnancy compared to women of similar age.[75] Estrogen is associated with a prothrombotic state and, like obesity, subject to the theory of microthrombotic obstruction of arachnoid villi, but there is no evidence supporting disruption of arachnoid villi.

Progesterone has mineralocorticoid effects which can enhance CSF production. Polycystic ovary syndrome is also known to occur more frequently in the PTCS population,[57] and this syndrome is associated with high levels of androgen. It is unlikely that the androgens themselves predispose to PTCS, but there is enzymatic conversion to estrogens peripherally that may be the significant factor. In addition to inducing retinoic acid synthesis, estrogen also decreases tight junctional resistance and remodeling of occludin which directly affects paracellular permeability.[117] Paracellular permeability is important in the exchange of water between the CSF and brain via the ependymal and pial lining. Rats given exogenous estrogen developed cerebral edema.[118] There is no suggestion that edema exists in PTCS, as pathologic studies have failed to identify brain edema.[119] However, it appears that the brain parenchyma is increased in size, as evidenced by the often smaller ventricular size and smaller venous sinus volume.[120] There are no volumetric studies on the brain as yet to prove this. Extra fluid taken into the interstitial fluid of the brain if taken up also by glial cells in balance will result in no edema being evident on imaging.

Estrogen may influence the exchange of water by influencing aquaporins. Although there is no evidence within the CNS, Carreras and colleagues demonstrated that expression of hepatocyte AQP8 water channels as well as canalicular membrane water permeability is down-regulated in estrogen-induced intrahepatic cholestasis.[121] Here, we have a link between estrogen and membrane permeability, and theoretically, estrogen could alter brain aquaporins. Overall, it can be concluded that estrogen, both endogenous and exogenous, may be associated with PTCS, although the extent of risk and the important mechanisms have not been defined.

Summary

The pathophysiology of the pseudotumor cerebri syndrome has been something of an enigma for a long time, with many suspected risk factors. The widespread use of noninvasive and minimally invasive venous imaging techniques has brought us back full circle to a vascular component, with frequent transverse sinus stenosis, but this is likely secondary to the increased pressure in most cases. Hypothesizing that primary venous sinus stenoses are etiologic does not appear very plausible given the literature available. What is more plausible is the stenosis might worsen the condition and set up a vicious “feed forward” cycle, relieved by the removal of CSF. Although aldosterone may be a common factor in many cases of IIH, associated with obesity and PCOS, enhanced activity of the Na/K-ATPase in the choroid plexus would increase CSF secretion, which is not evidenced by any studies to date. The lack of morphological changes in the choroid plexus, and the multiple outflow pathways for CSF, make this theory less likely. Multiple experimental studies have demonstrated increased resistance to CSF outflow in PTCS, and this appears the proximate mechanism. This resistance may be a field effect on epithelial and other membranes, reducing the flow of water. Leading contenders for this action are estrogen and possibly retinoic acid. Obesity may produce more of both. It is possible that pseudotumor cerebri does not occur in older age because of the low level of sex hormones and reduced CSF production. The increasing CSF pressure with reduced outflow in PTCS increases the water flowing into the interstitial fluid space of the brain, through the ependymal and pial surfaces, the cells of which are linked with gap junctions allowing paracellular flow. This extra water is taken up in proportion by glial cells also, and overall there is no edema, but the brain is mildly larger. As yet there is no volumetric evidence for this theory. There are likely genetic predispositions, which make the individual more susceptible to endogenous and exogenous risk factors for PTCS.

Statement of Authorship

Category 1

  • (a)Conception and DesignBrian E. McGeeney
  • (b)Acquisition of DataBrian E. McGeeney; Deborah I. Friedman
  • (c)Analysis and Interpretation of DataBrian E. McGeeney; Deborah I. Friedman

Category 2

  • (a)Drafting the ManuscriptBrian E. McGeeney; Deborah I. Friedman
  • (b)Revising It for Intellectual ContentBrian E. McGeeney; Deborah I. Friedman

Category 3

  • (a)Final Approval of the Completed ManuscriptBrian E. McGeeney; Deborah I. Friedman

Ancillary