Normal intracranial pressure

In normal circumstances intracranial pressure is maintained by cerebral arterial pressure which itself is subject to cerebral autoregulation such that, other things being equal, intracranial pressure remains constant over a wide range of systemic arterial blood pressure. Intracranial pressure is also greatly influenced by cerebral venous pressure. Furthermore, intracranial pressure is determined by CSF formation and absorption, but whether there are any physiological regulatory mechanisms operating at the choroid plexus or arachnoid villi and granulations is unclear. Pressure in CSF varies enormously in the lumbar region and at the vertex depending on posture (reviewed in Fishman11).

Increased intracranial pressure

At a simple level, various perturbations could lead to an increase in intracranial pressure without the development of hydrocephalus or florid visible abnormality on structural imaging. These are summarised in table 2. For any of these mechanisms to be operative, it is necessary that any compensatory processes are no longer functioning. Thus an increase in cerebral volume with an equivalent reduction in CSF volume will obviously not change the status quo. Over the years investigational techniques of every imaginable degree of complexity and invasiveness have been used to explore these possibilities in IIH. Many of the relevant indices such as CSF formation rate, CSF outflow resistance, CSF outflow rate, and sagittal sinus pressure can be measured or calculated, but some of the techniques used require certain assumptions and are therefore possibly fallible. Particular difficulties exist in knowing to what extent the brain is compressible in response to increasing CSF pressure, and to what extent the CSF space is expandable. These factors influence CSF outflow resistance calculations in infusion or perfusion studies.

Increased cerebral volume

Computed tomography offered a way of assessing cerebral volume in IIH, albeit somewhat crudely. A reduction in the size of the ventricular system, indicating an increase in cerebral volume, was reported in some studies,12-14 but not in others,15 and it remains controversial as to whether cerebral volume is significantly increased in IIH. The disagreement perhaps reflects heterogeneity of pathogenesis. The hope has been expressed that MRI will provide a great deal more information about what is going on in IIH, but thus far there has not been an abundance of reported studies of cerebral and CSF volumes in IIH, nor of the composition of cerebral tissue. Moser et al 16 reported an increase in white matter water signal, suggesting diffuse mild oedema, and Gideonet al 17 detected increased water mobility in subcortical white matter. Both studies required the use of special MRI sequences, routine sequences showing no abnormality. The brain in IIH has also been studied by positron emission tomography. Notably, no change in regional cerebral blood volume was found.18 The most invasive studies of cerebral tissue in IIH were cerebral biopsies, which were reported by Sahs and Joynt19 to show evidence of interstitial cerebral oedema, but necropsies20 have not confirmed those findings, nor did a review of some of the original biopsy material of Sahs and Joynt.20

Increased CSF production

Increased CSF production rate has been proposed as a mechanism of IIH. The production rate of CSF can be measured in patients, but the procedures (infusion or perfusion techniques) are invasive.21 In one study increased CSF production rate was reported in IIH.22 However, most investigators have not found CSF hypersecretion in IIH. An attempt at measuring CSF production rate non-invasively by recording CSF flow through the cerebral aqueduct using MRI gave highly variable results, but again did not support the view that CSF hypersecretion is important in IIH.23 The only condition in which the CSF production rate is known definitely to be increased is choroid plexus papilloma, a fairly rare paediatric tumour. With this tumour the situation can be complicated by obstructive hydrocephalus and hydrocephalus related to intraventricular haemorrhage. However, CSF overproduction has been proved in patients including one with a small non-obstructing tumour,24 25and it is presumptively part of the cause of the hydrocephalus. An IIH-like syndrome has not been reported in choroid plexus papilloma. Kollar and Johnson26 used embolisation to treat an arteriovenous malformation which involved the great vein of Galen in a 5 year old child. After the successful occlusion, which was not complicated by venous thrombosis, the patient developed a pseudotumour syndrome, and the proposal was that a result of the procedure might have been an increase in CSF production by the choroid plexuses of the lateral ventricles as a result of change in venous pressure in the choroid plexuses. The CSF production rate, however, was not measured. There has been speculation that the benign intracranial hypertension associated with hypervitaminosis A might be due to CSF hypersecretion, but evidence is lacking.27

Idiopathic intracranial hypertension would require a generalised increase in intracranial pressure without a significant pressure gradient across the cortical mantle, and without any capacity for the brain to be compressed. Mathematical modelling of ventricular size in the circumstance of increased CSF production predicts hydrocephalus, not IIH.28 Experimental infusion of artificial CSF into the lateral ventricles of dogs leads to modest ventricular enlargement, not an IIH-like syndrome.29

CSF outflow reduction

Much more important and relevant is the likelihood that impaired outflow of CSF into the venous system is a cause of IIH. However, herein lies a conundrum, as exactly the same mechanism is invoked to explain communicating hydrocephalus, including so called normal pressure hydrocephalus. In some studies the two groups of patients have even been analysed together, despite their striking clinical differences (for example, Borgesen and Gjerris30). An increase in CSF pressure, either due to CSF overproduction or due to impaired absorption, would be expected to lead to an increase in CSF volume, if the CSF space had the capacity for any expansion. Within a non-expansile skull and relatively non-expansile spinal canal, CSF could only easily accumulate at the expense of cerebral blood volume. In IIH there is neither a reduction in cerebral blood volume, nor an increase in CSF volume. In hydrocephalus, the main mechanism of ventricular dilatation is evidently a pressure difference between the ventricular CSF and the convexity CSF, but pressure atrophy is also thought to play a part in the ventricular dilatation (see Fishman11). However, pressure atrophy does not seem to be operative in patients with chronic IIH, except possibly for two patients reported by Malm et al 31 who developed hydrocephalus after years of IIH. If the proposition is that the impairment of outflow of CSF is a lesion at the arachnoid villi and granulations level, then there is no reason to expect any transmantle pressure gradient, and it is easier to envisage this as a mechanism for IIH than NPH. Infants might represent a special case, as a non-acute increase in intracranial pressure may be expected to cause expansion of the skull vault, allowing the accumulation of CSF, either inside the ventricles or outside (external hydrocephalus). However, in the mathematical model of Rekateet al,28 an increase in CSF outflow resistance alone leads to hydrocephalus, and to generate the conditions found in IIH a reduction in brain compressibility is required as well.

There is ample evidence from infusion and perfusion studies that IIH is associated with an impairment of outflow of CSF.31-36There is no direct evidence of dysfunction of arachnoid villi and granulations in IIH. Abnormalities of arachnoid villi have, however, been noted in certain conditions which involve raised intracranial pressure. Microscopy after subarachnoid haemorrhage has disclosed apparent obstruction of villi by cells and morphological changes in arachnoid villi and granulations.37 The outflow resistance of CSF is known to be increased in experimental subarachnoid haemorrhage.38 But the disturbance of CSF dynamics associated with subarachnoid haemorrhage is hydrocephalus, and the relevant site of CSF flow disturbance might be proximal to arachnoid villi and granulations. The same considerations apply to meningitis and experimental meningitis.39 However, a pseudotumour syndrome is sometimes seen in the context of meningitis (see for instance Cremer et al 40). Very high CSF protein concentration (spinal tumour, Guillain-Barré syndrome) is sometimes a cause of raised intracranial pressure with papilloedema,41 42 and the suggestion has been made that the protein leads directly to impaired CSF outflow, and there is experimental evidence to support this.43 Interestingly, some patients with raised intracranial pressure attributed to high CSF protein concentration from a spinal tumour develop hydrocephalus, and some have papilloedema without ventricular dilatation.44Agenesis, deficiency, or dysplasia of arachnoid villi and granulations leads to hydrocephalus in infancy.45 46 (As indicated above, the capacity of the infant skull to expand may explain why hydrocephalus rather than a pseudotumour syndrome develops in this situation.) Vitamin A deficiency can cause a pseudotumour syndrome. Morphological abnormalities of arachnoid villi and granulations in experimental vitamin A deficiency have been described, and are presumably the cause of the increased CSF outflow resistance and the raised intracranial pressure.47 Regrettably arachnoid villi and granulations were not available for histological examination in the two patients with IIH who came to necropsy and were reported by Wall et al.20

An apparent difficulty with the idea that IIH is caused by any sort of impairment of CSF outflow is the normal or even low CSF protein concentration in IIH.48 49 The fluid which is made by the choroid plexuses is principally water and salt, with a low protein concentration. Protein gets into CSF diffusely throughout the system either from the brain and spinal cord parenchyma (mainly getting there across the blood-brain barrier) or directly across the blood-CSF barrier. Protein is absorbed into the venous system along with CSF. The gradient of CSF protein concentration (low in ventricular fluid and higher in lumbar fluid) is thus easily understood. In addition, the permeability of the blood-CSF barrier may be greater in the lumbar region.11 An increase in CSF outflow resistance might be expected to involve an increase in CSF protein concentration, even if once a steady state is re-established the overall CSF turnover is unchanged. The low CSF protein concentrations sometimes found in IIH are measured in lumbar CSF, and a possible way of accounting for this would be an increase in CSF absorption at a spinal level in the face of an impairment of CSF absorption into the superior sagittal sinus (a proportion of CSF is absorbed by arachnoid villi and granulations which are in veins around spinal nerve roots). Low lumbar spinal fluid protein concentration would be compatible with a lesion affecting cerebral arachnoid villi and granulations selectively or preferentially. It would also be compatible with an increase in cerebral venous sinus pressure, but presumably not a global systemic increase in venous pressure which would affect spinal as well as cranial CSF absorption. Furthermore it would be compatible with an inverse relation between CSF protein concentration and intracranial hypertension. Such a relation was reported by Chandraet al,49 but not confirmed by Johnston et al 48in a larger study with more robust data.

Intracranial venous hypertension

The final candidate mechanism for IIH is the obvious one of an increase in venous sinus pressure—obvious because lesions which increase venous sinus pressure (for example, dural arteriovenous malformations) or impede venous drainage (for example, venous sinus thrombosis, malignant obstruction of venous sinuses or jugular veins) are known to give rise to the same syndrome as IIH.50-53Clearly superior sagittal sinus thrombosis will affect cerebral venous pressure and drainage and will also directly affect CSF absorption, but any disorder causing a rise in venous pressure will secondarily have an effect on CSF absorption.

In the CT era it is in fact quite likely that cases of cerebral venous sinus thrombosis were misdiagnosed as having IIH, as the diagnosis was often made on the basis of the clinical picture, an unremarkable scan and a lumbar puncture. Magnetic resonance imaging and magnetic resonance venography have improved the reliability of non-invasive detection of cerebral venous sinus thrombosis, but still some cases may be missed without catheter angiography or venography.54 55 Recent reports of potentially prothrombotic abnormalities of coagulation in IIH56-58may be construed as indicating that undetected cerebral venous sinus thrombosis remains a mechanism of IIH, although other interpretations are possible.

Different groups have proposed that increased intracranial venous pressure is the major mechanism of raised intracranial pressure in IIH.31 54 59 60 Malmet al 31 reported a long term study in which patients with IIH underwent repeated assessments of CSF hydrodynamics by means of a constant pressure infusion technique. In most of their patients raised CSF pressure could be explained by increased sagittal sinus pressure. Their hypothesis was that the increase in pressure in the superior sagittal sinus was secondary to cerebral swelling leading to a reduction of the diameter of the superior sagittal sinus, but as has been pointed out above brain swelling is not necessarily seen in IIH. Their other patients had raised pressure on the basis of reduced CSF outflow conductance, presumed to reflect a lesion at the arachnoid villi and granulations level.

In nine patients studied by King et al,54 little abnormality was visible in the venous phase of cerebral angiograms, but manometry documented raised pressures in the superior sagittal sinuses and proximal transverse sinuses, with a drop in pressure in the distal transverse sinuses. Venography showed narrowing of the transverse sinuses, with either smooth tapering of uncertain cause, or intraluminal filling defects suggestive of mural thrombus. Of note were two patients whose intracranial hypertension was attributed to minocycline, who did not have raised venous sinus pressures, suggesting heterogeneity of pathogenesis. Subsequently Kinget al 61 reported briefly on a larger patient series. Fifteen out of 17 patients with IIH had raised superior sagittal sinus and proximal transverse sinus pressures with a drop in pressure in the distal transverse sinus. In four of these patients CSF was removed at the time of manometry with a resultant lowering of intracranial pressure, and that led to abolition of the apparent functional obstruction of the distal transverse sinus, which suggested to the authors that intracranial hypertension caused compression of the transverse sinus in some patients. This study highlights the possibility that increases in CSF pressure and venous pressure can interact so that each makes the other worse. The authors imply that they do not consider the increase in venous sinus pressure to be the primary event in most of their patients.

By contrast, Karahalios et al 59speculated that “most if not all aetiologies (of IIH) result in an increase in intracranial venous pressure as a final common pathway.” In their series venous outflow obstruction was detected by venography in five out of 10 patients studied. In the remaining five there was no obstruction but venous pressures were nevertheless increased, as were right atrial pressures with transmission of the raised central venous pressures back to the intracranial venous system. Karahalioset al 59 discuss ways in which obesity might lead to raised central venous pressures, but conclude that the mechanism of increased central venous pressure in IIH remains obscure.

No such doubts in the mind of Sugarman et al,60 62 63 who contend that at least in morbidly obese persons pseudotumour cerebri is a direct result of obesity which leads to increased central venous and intracranial pressures (see below). Indeed they maintain that intracranial hypertension in this situation should no longer be considered idiopathic. In their hands gastric bypass surgery had a high success rate in resolution of symptoms of raised intracranial pressure (as well as treating joint problems, gastro-oesophageal reflux, high blood pressure, sleep apnoea, hypoventilation, diabetes, and urinary incontinence!). It might be speculated that increased venous pressure would affect cerebral compressibility in such a way as to favour a pseudotumour syndrome rather than hydrocephalus. Unexplained by this hypothesis is the absence of pseudotumour syndrome as a complication of right ventricular cardiac failure, although increased CSF pressure has been shown to accompany the rise of venous pressure which occurs in right heart failure.64

Obesity and IIH

The relation between IIH and obesity has long been recognised.65-67 Pressure in the CSF is higher in obese but otherwise normal people than in people of normal weight.68 An association between recent weight gain and the development of IIH has been established.65 66 Weight reduction has long been part of the treatment strategy. There is some evidence that weight reduction is therapeutic. In two retrospective studies, Kupersmith et al 69 and Johnson et al 2 independently found that weight reduction was associated with improvement in69 or resolution of2 papilloedema in their patients. Rowe and Sarkies70 on the other hand found no correlation between weight change and visual improvement in their series. Sugarman et al 60reported on a series of eight morbidly obese patients with IIH, in all of whom gastric surgery for weight reduction was successful in bringing about considerable weight reduction and also resolution of symptoms and signs of IIH including reduction of CSF pressure to normal.

How does obesity lead to an increase in intracranial pressure? Hormonal mechanisms have been postulated,70 but Sugarmanet al 62 have provided persuasive evidence in morbidly obese people for a simpler mechanism. They showed that their obese patients with IIH had raised intra-abdominal pressures, raised intrathoracic pressures, and raised central venous pressures, supporting a direct cause and effect relation. However, they did not have a satisfactory, concurrent control group, and did not provide an explanation as to why the pulmonary artery pressures were higher in their obese patients with IIH than in equivalently obese patients without IIH. It would be easy to diagnose this patient group as just having a pickwickian syndrome, but only two out of eight morbidly obese patients of Sugarman et al 60 had evidence of alveolar hypoventilation with reduced blood oxygen and raised blood carbon dioxide concentrations. Obesity is common, whereas IIH is rare, and obesity affects males as well as females, whereas IIH is much commoner in females than males, so obesity cannot be a sole cause of IIH, even though it may be the immediate cause in a suitably predisposed person. The established relation between obesity and IIH raises further questions. Although IIH in obese persons may be rare, headache in people of all shapes and sizes is very common, and the question does arise as to whether a proportion of obese persons with headache but without papilloedema in fact do have IIH. In the classic IIH series, papilloedema was present in 100% cases and was a diagnostic sine qua non.71-74However, it is not necessary to modify the Dandy criteria much further to open up a whole new vista of cases of overweight patients with bad headache and with high CSF pressures. Indeed, only 12 out of 24 patients had papilloedema in the updated series of Sugarmanet al.63 Several centres have published on IIH without papilloedema.75-77Disappointingly, the response of the patients to medication aimed at lowering intracranial pressure was poor.

Conclusion

It seems inescapable that the condition currently called IIH is heterogeneous, and indeed Johnston et al 78 proposed using the term pseudotumour syndrome to encompass this heterogeneity. In some patients there may be just one aetiology operating, such as occult venous sinus thrombosis. Perhaps others have risk factors which combine to precipitate the condition. In particular, in many cases obesity may be a risk factor whereas in extreme cases it may be a sufficient cause. Brain MRI really should be able to provide definitive information about cerebral and CSF volumes in IIH, but as yet the tunnel has not shed much light, and intriguing enigmas remain. Laboratory animal research, if possible, into factors influencing function of the arachnoid villi might well be informative. More effective means of preventing or treating obesity would undoubtedly have an impact on the prevention and treatment of IIH.