Elevated intracranial venous pressure as a universal mechanism in pseudotumor cerebri of varying etiologies

There have been several theories suggested to explain the etiology of pseudotumor cerebri (PTC). Dandy ^[1] hypothesized that the volume of either cerebral blood or CSF might be increased in this condition. Similarly, Foley ^[2] described increased cerebral blood flow in association with elevated intracranial pressure (ICP). Sahs and Joynt ^[3] reported microscopic evidence of "intracellular and extracellular edema," and Raichle et al ^[4] suggested that PTC may be caused by an abnormality in the cerebral microvasculature, leading to an increase in the water content of tissue. Johnston et al ^[5,6] suggested that the underlying mechanism in PTC involves disturbed CSF absorption secondary to increased sagittal sinus pressure, which reverses the CSF-to-sinus pressure gradient necessary to drive bulk flow of CSF. Malm et al ^[7] concluded that increased CSF pressure in PTC is explained by two mechanisms: a rise in sagittal sinus pressure, which in turn is the result of compression of the sinus by extracellular edema, or a decreased conductance with a compensatory increase in CSF pressure to sustain bulk flow.

We undertook the present study to determine whether a common variable could be identified in idiopathic PTC and in PTC associated with other conditions, such as intracranial venous outlet obstruction. Our previous clinical experience suggested that increased venous pressure may be a common finding in PTC of various etiologies. We therefore tested this hypothesis using current endovascular techniques.

Methods.

Ten patients (three males, seven females; age range, 2 to 40 years) with signs and symptoms consistent with the diagnosis of PTC, as defined by modified Dandy criteria, ^[1,8,9] presented to the neurosurgical service after more conservative means of treatment had failed. CT or MRI was performed in all patients without previous imaging studies to confirm the absence of intracranial mass lesions.

CSF pressure was measured in each patient by lumbar puncture, Richmond subarachnoid bolt, or a Camino fiberoptic ICP monitoring device (Camino Laboratories, San Diego, CA). Lumbar punctures were performed with the patients in the lateral decubitus position, with legs extended during pressure measurements. Intracranial ICP monitors were placed in the right frontal region, and pressure measurements were made with the patient's head elevated to 30 degrees.

In patients undergoing MRI, MR venograms were also performed. All ten patients underwent conventional angiography (venography) via the femoral vein approach. Intracranial venous (dural sinus) and central systemic venous (right atrial) pressures were measured during angiography, when possible. In some patients with venous sinus outflow obstruction, angioplasty, thrombolytic therapy, or both were undertaken.

For angiography, patients were sedated with intravenous analgesia. The right groin was shaved and then prepared with an antiseptic povidone-iodine-based solution. Lidocaine was used for local anesthesia. An 18-gauge needle was used to puncture the right femoral vein. A 0.035-inch J-wire was used to introduce a 5.5-French-diameter headhunter catheter. A Tracker 18 microcatheter with a Seeker 14 guide wire (Target Therapeutics, Fremont, CA) was introduced coaxially using a Y-valve connector and navigated under fluoroscopic conditions and roadmapped to the various regions of interest within the cerebral venous circulation and the heart. Pressure was measured using a transducer connected to the Tracker 18 microcatheter. The microcatheter was also used to perform the venous angiography.

If thrombolysis was attempted, the microcatheter was left in contact with the clot, and urokinase or tissue plasminogen activator (tPA) was infused. The patients then underwent venous angiography again to determine whether the obstruction had resolved. These patients also underwent systemic anticoagulation with heparin with a partial thromboplastin time goal between 35 and 45 seconds.

If no thrombus was present or if thrombolysis was ineffective, angioplasty was attempted. A Seeker 14 guide wire was navigated across the area of stenosis or residual obstruction, and a Stealth balloon (1 cm long x 3 mm in diameter; Target Therapeutics) was passed across. Under roadmap conditions, the balloon was maneuvered across the area of stenosis and inflated. If necessary, balloon inflation was repeated until the obstruction was eliminated.

A positive response to treatment was defined as either an improvement in symptomatology, a resolution of papilledema, or a decrease in CSF pressure/ICP. All 10 patients underwent surgical treatment, including lumboperitoneal shunting (n = 10), optic nerve sheath fenestration (n = 3), or gastric stapling (n = 1).

Results.

Presenting symptoms included headache (n = 7), visual difficulties (n = 3), and scalp venous distention (n = 1). One patient was asymptomatic except for papilledema. On physical examination, most patients had some degree of papilledema (n = 7). Other findings included decreased visual acuity (n = 2), sixth-nerve palsy (n = 1), and dilated scalp veins (n = 1). One patient had a normal examination despite a complaint of severe headache.

On radiographic studies, no patient had a mass lesion that could account for the increased ICP/CSF pressure. The brain parenchyma appeared normal in all 10 patients, and there was no evidence of hydrocephalus. One patient, with Camurati-Engelmann syndrome, had severe bony overgrowth of the skull.

CSF pressure was elevated in all 10 patients (>200 mm CSF). Of the five patients who underwent MR venography, three had evidence of intracranial obstruction of venous outflow. Of the two remaining patients with negative MR venograms, one had venous outflow obstruction demonstrated by conventional venography and the other had normal venous sinus anatomy and no evidence of obstruction. Conventional venography demonstrated venous sinus outflow obstruction in five patients, secondary to either thrombosis (n = 2), tumor (n = 1), or congenital stenosis (n = 2) (Table 1).

Thrombolysis with infusion of urokinase or tPA, attempted in two patients, failed to resolve the thrombosis. Angioplasty was attempted in three patients. In two patients the stenosis improved. One of these patients, however, suffered symptomatic restenosis one year later, and the other patient continued to have asymptomatic papilledema despite an excellent hemodynamic result. In one of these patients the pressure gradient across the stenosis decreased from 12 to 4 mm Hg, and from 75 to 10 mm Hg in the second.

Dural sinus pressure was measured during venography in eight patients (Table 1 and Figure 1). Normal central venous pressure measured at the right atrium in the supine adult ranges from 0 to 4 mm Hg, ^[10] and normal superior sagittal sinus pressures in the supine adult from 4 to 10 mm Hg. ^[11,12] In our series of patients with PTC, pressure in the superior sagittal sinus ranged from 13 to 24 mm Hg (mean, 16.6 mm Hg). Patients with obstruction tended to have a high pressure gradient across the obstruction, with high pressure proximally and lower pressure distally. Five patients with normal venous anatomy (no obstruction) had elevated right atrial pressures (range, 6 to 22 mm Hg; mean, 11.8 mm Hg) as well as elevated venous sinus pressures. Right atrial pressure measured in one of the five patients with intracranial venous outflow obstruction was within normal limits (4 mm Hg).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Diagrammatic presentation of the dural venous sinuses. Posted pressures are in mm Hg and represent the mean pressures for each group at each anatomic location. (A) Pseudotumor with venous outflow obstruction (n = 5). (B) Pseudotumor without outflow obstruction (n = 5). Reprinted with permission of Barrow Neurological Institute.

All 10 patients underwent surgical treatment for their disease. Six patients were treated with lumboperitoneal shunting, three with optic nerve sheath fenestration, and one with gastric stapling. Follow-up ranged from 4 to 8 months. All patients improved symptomatically, and the papilledema resolved in all patients who had initially presented with this symptom.

Discussion.

The present study strengthens the theory that PTC, whether secondary to intracranial venous outlet obstruction or idiopathic, may have increased venous pressure as its final common pathway. In patients with venous outflow obstruction, as demonstrated by this study, the resistance to venous drainage generates the elevated venous back pressure. Patients with idiopathic PTC (without outflow obstruction) had elevated central systemic venous pressure (right atrial pressures), which may be transmitted up through the jugular veins and paravertebral plexus to the sinuses. This contrasts with the conclusions of King et al, ^[13] who also demonstrated increased dural venous pressure in patients with "idiopathic" PTC but felt that it was secondary to partial outflow obstruction. As we and others suggested previously, ^[5,6,14-23] the elevated sinus pressure increases the resistance to CSF absorption. CSF pressure subsequently increases to restore the pressure gradient necessary for CSF to flow across the arachnoid villi and into the venous system. CSF secretion likely continues despite the increasing ICP, further exacerbating intracranial hypertension. CSF production does not appear to increase, however, and therefore cannot account for the increase in ICP. ^[6,14,15,24] Increased venous pressure and the resultant increased blood volume may also contribute to the increase in ICP, ^[1] but this component likely plays a minor role. ^[4]

While PTC patients with outflow obstruction may be able to absorb CSF through alternative higher-resistance channels such as the spinal arachnoid villi, idiopathic PTC patients with central systemic increases in venous pressure may not be able to drain CSF through alternative channels. If so, idiopathic PTC patients would be expected to have higher ICP than patients with outflow obstruction, but we did not find this in our series. Furthermore, increased ICP from increased CSF pressure may expand CSF volume modestly. This increased volume is limited by the rigid cranial and spinal compartments. The subarachnoid space, however, can expand into the spinal nerve root sleeves; into the optic nerve sheaths, causing papilledema; and into the sella turcica, causing the empty-sella syndrome. ^[25,26] Others have suggested that increased parenchymal water content or frank edema may account for expansion of CSF volume. ^[3,4,7] The issue of parenchymal swelling in PTC continues to be controversial, as is evident by the recent paper of Wall et al ^[27] that challenged the brain edema model of Sahs and Joynt. ^[3]

Increased intracranial venous pressure is also linked to the development of hydrocephalus. ^{[16-18,20,28-30]} Experimental obstruction of venous outflow leads to cerebral venous hypertension and produces hydrocephalus in dogs. ^[20] Unlike in young children with open sutures, intracranial venous hypertension and the resultant increase in ICP may not lead to hydrocephalus when the cranial sutures are closed. ^[18,30] In the present study, all patients had closed sutures (ie, a nondistensible cranial compartment) and thus developed PTC instead of hydrocephalus. One child with a venous outlet obstruction initially presented with hydrocephalus before his sutures closed. A shunt was placed and led to normalization of his ventricular size. Later, however, when his sutures were closed, he presented with PTC (normal-sized ventricles).

This reasoning is consistent with our previously published mathematical model for ventricular volume regulation. ^[21] However, a second effect of increased resistance to venous outflow must be considered. As resistance to venous drainage increases, brain turgor, or K_b, also increases. ^[22] The stiffened brain resists compression as predicted by the mathe-matical model and subsequently leads to PTC. Conversely, in normal-pressure hydrocephalus, CSF flow is restricted between the spinal and cortical subarachnoid spaces, leading to a more compressible brain (lower K_b). Ventricular volume thus increases with little increase in ICP.

The etiology of elevated central systemic venous pressure in idiopathic PTC is unknown. In the present study, the group with idiopathic PTC was composed of obese women between the ages of 25 and forty. This profile typifies the patient who presents with idiopathic PTC. The precise relationship between this profile and elevated central venous pressure is unknown. However, obesity could be related to the increase in central systemic venous pressure by a number of mechanisms (Figure 2). Obesity-related cardiomyopathy can lead to congestive heart failure and a subsequent increase in central venous pressure. ^[31] Obesity also can lead to sleep apnea or increased work in breathing, which in turn leads to respiratory acidosis, right-sided heart failure, and thus increased central venous pressure. ^[32,33] Sugerman et al ^[34] have suggested that carbon dioxide retention in these patients may lead to increased ICP. They also explained that the increased systemic venous pressure may be secondary to an increase in intra-abdominal pressure, which in turn leads to increased pleural pressure and decreased venous return from the brain to the heart. This also is consistent with our right-sided heart failure model. The usual physiologic mechanisms for correcting increased plasma volume also may be aberrant in obesity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Flow diagram demonstrating the various mechanisms by which obesity can lead to increased central systemic venous pressure (CVP). Shaded area represents normal physiology. CHF = congestive heart failure; ANF = atrial natriuretic factor; HTN = hypertension; DM = diabetes mellitus; Na sup + = sodium.

In response to increases in central venous pressure and the subsequent distension of the right atrium, atrial natriuretic factor (ANF) is released. ^[35] In turn, ANF acts on the kidney to facilitate the excretion of sodium and water in the normal state. ^[36] However, in certain conditions, such as in response to heart failure, the release of ANF may not lead to sodium and water excretion because the kidney is refractory. ^[37] In the obese patient, associated conditions such as diabetes mellitus, hypertension, and nephrotic syndrome may limit the kidney's ability to excrete salt and water, which again may be secondary to frank organ failure or to the kidney's refractoriness to ANF. ^[38,39] Furthermore, an increase in central venous pressure and thus in renal venous pressure may lead to sodium and water retention by a direct action on the kidney. ^[40] The finding of peripheral edema in our patients with idiopathic PTC also may be explained in terms of these pathologic pulmonary, cardiovascular, and renal mechanisms. ^[33] The roles of these mechanisms in increasing central venous pressure in idiopathic PTC have yet to be demonstrated. It is clear, however, that idiopathic PTC, which is commonly found in young obese females, is a systemic disease.

Patients, both with and without obstructions, improved when treated with conventional CSF diversion procedures, suggesting that this approach remains an effective treatment for PTC patients who are refractory to the usual medical therapies. Treatment strategies aimed at addressing the elevated venous pressures tend to be less reliable and practical than the conventional CSF diversion procedures. With respect to PTC caused by venous outflow obstruction, our present study suggests that endovascular techniques used to open areas of stenosis can lead to angiographic improvement and, in some cases, to a decrease in pressure gradients and symptoms, but not consistently and reliably. Vascular bypass of a venous sinus improves hydrocephalus and decreases ICP, ^[18] but this is a high-risk procedure prone to graft occlusion.

Medical treatment is usually effective in patients with idiopathic PTC. Diuretics may exert their primary effect by decreasing central plasma volume, and hence venous pressure, as well as by decreasing CSF production. In patients with idiopathic PTC refractory to medical treatment, CSF diversion is effective. Alternatively, in obese patients undergoing gastric stapling, the subsequent weight loss may reverse the strain placed on the heart and other organs, thereby resolving the occult failure and central venous hypertension. Sugerman et al ^[34] have demonstrated a dramatic decrease in CSF pressure and symptoms in obese patients treated with gastric stapling.

This study suggests that elevated intracranial venous pressure may be a universal mechanism in PTC of various etiologies. A prospective study is needed to demonstrate that elevated central systemic and intracranial venous pressure exists in PTC associated with other, more obscure clinical conditions, such as endocrine or nutritional disorders, or with the use of certain drugs. ^[41] If this variable is indeed the final common pathway of PTC, it may be possible to formulate more effective treatment strategies that may prove to be more reliable and fraught with fewer complications than decreasing ICP by diverting CSF.