INTRODUCTION — The production, absorption, and flow of cerebrospinal fluid (CSF) play key roles in the dynamics of intracranial pressure. Alterations in CSF pressure can lead to neurologic symptoms, the most common being headache. Most often, the headaches associated with low CSF pressure are orthostatic and occur after lumbar puncture, but similar headaches occur with spontaneous low CSF pressure due to spinal CSF leaks, and with CSF shunt overdrainage [1,2].

Spontaneous intracranial hypotension is being recognized with increasing frequency. Orthostatic headache, low CSF pressure, and diffuse meningeal enhancement on brain magnetic resonance imaging (MRI) are the major features of the classic syndrome. However, some cases have nonorthostatic headache, normal CSF pressure, or no evidence of diffuse meningeal enhancement [2]. As our understanding of this syndrome is still evolving, the terms "low CSF pressure headache," "spontaneous intracranial hypotension," "spontaneous CSF leak," "CSF hypovolemia," and "CSF volume depletion" tend to be used interchangeably.

This topic will review the treatment and prognosis of spontaneous intracranial hypotension. Other aspects of this disorder are discussed separately. (See "Spontaneous intracranial hypotension: Pathophysiology, clinical features, and diagnosis".)

Post-dural puncture headache is reviewed elsewhere. (See "Post dural puncture headache".)

DIAGNOSIS — The diagnosis of spontaneous intracranial hypotension is reviewed here briefly and discussed in detail separately. (See "Spontaneous intracranial hypotension: Pathophysiology, clinical features, and diagnosis", section on 'Evaluation and diagnosis'.)

The diagnosis should be considered in patients who present with positional orthostatic headache, with or without associated symptoms, perhaps in the setting of minor trauma, and in the absence of a history of dural puncture or other cause of cerebrospinal fluid (CSF) fistula. Confirmation of the diagnosis (algorithm 1) requires evidence of low CSF pressure, usually by brain magnetic resonance imaging (MRI) (image 1 and image 2), and/or evidence of CSF leakage on other neuroimaging studies, mainly computed tomographic (CT) myelography.

TREATMENT — Epidural blood patching is the mainstay of treatment for spontaneous intracranial hypotension [3-5]. Patients with acute uncomplicated headache of mild to moderate severity can be treated initially with conservative therapy (algorithm 1). However, patients who present with headache that is severe or longstanding (ie, two weeks or more since onset) should be treated initially with an epidural blood patch (EBP), as should patients who have associated symptoms that are disabling or related to distortion or compression of brain and spinal cord structures (eg, diminished level of consciousness). (See "Spontaneous intracranial hypotension: Pathophysiology, clinical features, and diagnosis", section on 'Associated symptoms and complications'.)

Unfortunately, no randomized clinical trials have evaluated the effectiveness of the various treatment strategies employed, and no definitive treatment protocols have been established. Clinical experience suggests that some patients with this condition improve without specific treatment, given enough time. However, in many instances intervention is necessary, not only to speed recovery, but to reach full recovery.

For patients who fail adequate trials of conservative therapy and repeated EBP treatments, additional options include continuous epidural saline infusion, epidural fibrin glue, or surgical repair of the defect. Both epidural fibrin glue and surgical repair require definitive localization of the cerebrospinal fluid (CSF) leak or leaks. This is usually accomplished with CT myelography or, less often, radioisotope cisternography.

Conservative treatment — The most conservative treatment for spontaneous intracranial hypotension is avoidance of the upright position, with strict bed rest and the possible addition of analgesics. However, many patients are unable to comply with strict bed rest, particularly as its benefits can be very short-lasting.

Strategies aimed at restoring CSF volume include oral or intravenous hydration, high oral caffeine intake, and high salt intake. Use of an abdominal binder is an additional measure that may be helpful [6].

While analgesics are often recommended as first-line treatment, generally, they provide little relief [7]. Anecdotally, glucocorticoids have been reported to be of some benefit, but remain unproven [3].

For patients with acute, uncomplicated mild to moderate spontaneous intracranial hypotension, we suggest use of conservative measures as initial therapy. Our preferred regimen consists of bed rest and generous caffeine intake. Approximately 200 to 300 mg of caffeine, given two to three times daily, generally provides short-term relief. Caffeinated beverages (eg, coffee or soda) or caffeine tablets can be used. With improvement, patients can return to activity gradually, incorporating scheduled rest periods lying flat (eg, five minutes per hour).

Oral hydration is traditionally recommended but its effectiveness has not been demonstrated in clinical studies [8]. However, by the time that many patients with mild to moderate spontaneous intracranial hypotension present to a neurologist, they have already had the headache for weeks to months, and conservative measures have failed. For those patients, EBP is recommended as soon as possible.

Epidural blood patch — We suggest EBP as first-line therapy for patients with spontaneous intracranial hypotension who fulfill any of the following conditions [4,9]:

●Acute, mild to moderate headache unresponsive to a reasonable period of conservative treatment (eg, one to two weeks)

●Severe headache or other disabling symptoms, regardless of duration

●Symptomatic for two weeks or longer at the time of diagnosis

●An aggressive precipitating injury (eg, a water skiing accident) as compared with a minor or "trivial" trauma (eg, a sudden twist or stretch)

●A history of connective tissue disease or joint hypermobility

The mechanism that accounts for the success of EBP is not completely understood, but the leading hypothesis is that EBP works initially through tamponade of the dural leak, and later by fibrin deposition and scar formation that is found within three weeks [6]. The tamponade of the leak often results in an immediate improvement in symptoms.

Few studies have investigated the effectiveness of EBP for spontaneous intracranial hypotension, and the available evidence comes from small retrospective case series. The range of findings is illustrated by the following reports:

●A case series of 25 consecutive patients with spontaneous CSF leaks found a good response to the initial EBP in 9 (36 percent) of 25 patients and a good response to a second EBP in 5 (33 percent) of 15 patients [10]

●In a series of 30 patients diagnosed with spontaneous intracranial hypotension based upon clinical features, pain resolved completely with the first EBP in 17 patients (57 percent) and with a second EBP in 6 of 9 patients [11]

These studies highlight the observation that approximately 50 percent of patients with spontaneous intracranial hypotension require more than one EBP treatment, and some patients may require up to four to six patches [3,12]. In one study of 150 patients, features associated with a lack of response to a first targeted EBP included the length of anterior epidural CSF collection, severity of diencephalic-mesencephalic deformity, and an injected blood volume ≤22.5 mL [13]. Lumbar placement of the EBP can be effective even when the site of CSF leakage is above the site of the blood patch or is unknown. One study found that targeting the EBP to the level of the CSF leak (when known) was more effective than blind EBP at the lumbar level, but this finding is limited by methodologic limitations including the retrospective, nonrandom study design and small patient numbers [14].

The efficacy of EBP therapy is generally lower for spontaneous intracranial hypotension than for post-dural puncture headache, where success rates can be as high as 90 percent. In a review of 105 patients with symptomatic intracranial hypotension, patients with spontaneous intracranial hypotension required EBP more often than those with post-dural puncture headache (70.5 versus 45.5 percent) and typically more than once (37.7 versus 13.6 percent) [15]. While there is no single explanation for this, several factors may play a role. A lumbar puncture is a site-known "nick" in the posterior aspect of the dura. In contrast, a spontaneous CSF leak site is often site-unknown and is commonly above the lumbar region near the nerve root sleeve and thus distant from the EBP placement. These factors probably contribute to a reduced likelihood of success with the first EBP. Additionally, a spontaneous CSF leak is not a well-placed hole but rather a dural rent, tear, or connective tissue laxity. Therefore, the healing and "sealing" process following EBP placement is more complex and potentially more likely to leak again or never fully seal in the first place. Consequently, multiple EBP treatments often are required before success is achieved. (See "Post dural puncture headache", section on 'Epidural blood patch'.)

EBP regimen and side effects — The initial EBP treatment typically involves the infusion of 10 to 20 cc of autologous blood into the epidural space [6]. A larger volume (20 to 100 mL) infusion may be used if the initial blood patch is unsuccessful [16]. With use of the higher volume infusion, a minimum of five days between blood patches is suggested because of the potential for an increased risk of complications including back pain and radiculopathy. In addition, some experts advise the use of different injection sites for repeated blood patches (eg, first at the thoracolumbar junction, then in the lower lumbar area) [16]. An EBP directed at the cervical or thoracic spine may be preferred if leak site is strongly suspected to be higher; examples include CSF leaks related to athletic activity (eg, swimming, bench pressing, or playing tennis).

Adverse effects of EBP are uncommon but include back pain, radiculopathy, leg paresthesias, and fever. Transient bilateral paraplegia and cauda equina syndrome from arachnoiditis have been reported with high-volume EBPs [17].

Epidural fibrin glue — Epidural patching with fibrin glue at the site of the CSF leak has been used successfully in small numbers of patients [18-20]. Anecdotal evidence suggests that this percutaneous method is effective, and thereby avoids surgery, in about one-third of patients who have failed EBP treatment [16].

Further evidence of benefit in larger studies is needed before this technique can be routinely recommended [18]. Nevertheless, some authorities favor the use of epidural fibrin glue in an effort to avoid surgery for patients with a clearly identified site of CSF leak who have failed an adequate trial of repeated EBP.

Some commercially available fibrin sealant products contain bovine aprotinin, an antifibrinolytic agent. (See "Fibrin sealants", section on 'Components'.)

The use of fibrin sealant is associated with a very low risk of anaphylaxis. However, repeated treatment with fibrin sealant containing bovine aprotinin appears to be associated with a substantially increased risk of hypersensitivity reaction [21]. Thus, patients selected for epidural fibrin glue treatment should be asked about possible previous exposure to aprotinin or fibrin sealant, since these agents have been used for hemostasis in various types of surgery (eg, cardiac, transplantation, joint replacement, repair of splenic injury) and hemostatic disorders [22].

Measures that may reduce the risk of anaphylaxis, although unproven, include treating multiple sites of CSF leak together rather than as staged procedures, waiting three to six months before repeating fibrin sealant treatment, and performing repeat treatment after prophylactic administration of glucocorticoids, H1 receptor antagonists, and hospital admission for observation [21].

Surgical repair — Surgical repair of a CSF leak may be required in some patients if nonsurgical therapies fail [16,23,24].

Surgical methods involve the use of suture or metallic clips to ligate leaking meningeal diverticula [1,16,23]. Dural tears and defects can be repaired with suture or with placement of a muscle pledget, Gelfoam, and/or fibrin glue.

We suggest surgical repair only for patients with a clearly identified site of CSF leakage by neuroimaging (typically CT myelography or digital subtraction myelography) who have failed an adequate trial of repeated EBP. In one report of 14 patients with intractable symptoms, exploratory microsurgery allowed visualization and treatment of CSF leaks originating from longitudinal dural slits located on the ventral (n = 10), lateral (n = 3), or dorsal (n = 1) aspects of the dura [24]. All 10 of the ventral slits were caused by calcified microspurs that protruded from the intervertebral disc. However, an earlier report found that surgical exploration had a low yield for detecting the source of a CSF leak [23]. We suggest obtaining surgical treatment from surgeons or centers with expertise in spinal surgery for spontaneous intracranial hypotension.

Continuous epidural infusion — One method of restoring the intracranial CSF volume, and thereby reducing headache, is a continuous epidural infusion of saline or dextran [25]. This method has limited success, but it is suggested for patients who fail EBP if the site of the CSF leak cannot be identified. In addition, it may be useful as a temporizing measure when urgent treatment is needed (eg, because of stupor or coma) for those who fail EBP and are awaiting permanent repair of a CSF leak [16,26,27].

PROGNOSIS — Spontaneous intracranial hypotension may resolve spontaneously within two weeks [28]. In some cases, it may last months, or in rare cases, years. However, intermittent headaches have been reported at intervals of weeks, months, or years, probably caused by intermittent cerebrospinal fluid (CSF) leaks [16]. Furthermore, some patients have persistent symptoms despite documented resolution of CSF leakage with therapy [29].

Although only anecdotal data are available, spontaneous spinal CSF leakage is estimated to recur in approximately 10 percent of patients regardless of treatment [16].

Preliminary data suggest that initial brain magnetic resonance imaging (MRI) findings may be helpful for predicting the outcome of spontaneous intracranial hypotension. In a referral center series of 33 patients with spontaneous spinal CSF leaks and intracranial hypotension, a good outcome was reported in 25 (97 percent) of 26 patients with an abnormal MRI versus 1 (14 percent) of 7 patients with a normal MRI, and the difference between these groups was statistically significant [29]. However, it is unclear why a normal MRI was associated with a worse prognosis, and larger studies in more diverse patient populations are needed to confirm this finding.

SUMMARY AND RECOMMENDATIONS

●Conservative treatment – For patients with acute, uncomplicated spontaneous intracranial hypotension of mild to moderate severity, we suggest use of conservative measures as initial therapy (algorithm 1) (Grade 2C). Our preferred regimen consists of bed rest and generous caffeine intake. (See 'Conservative treatment' above.)

●Epidural blood patch – We suggest epidural blood patch (EBP) for patients with spontaneous intracranial hypotension who fulfill any of the following conditions (Grade 2C) (see 'Epidural blood patch' above):

•Headache unresponsive to a reasonable period of conservative treatment (eg, one to two weeks)

•Severe headache or other disabling symptoms, regardless of duration

•Symptomatic for two weeks or longer at the time of diagnosis

•An aggressive precipitating injury, a history of connective tissue disease, or joint hypermobility

●Options for refractory patients – For patients who fail adequate trials of conservative therapy and repeated EBP treatments, other interventions may be necessary to speed recovery and to reach full recovery.

•For patients with spontaneous intracranial hypotension who have failed an adequate trial of repeated EBP and have a clearly identified site of CSF leakage, we suggest surgical repair (Grade 2C). (See 'Surgical repair' above.)

•Anecdotal evidence suggests that treatment with epidural fibrin glue is beneficial for spontaneous intracranial hypotension. However, larger studies are needed before this technique can be routinely recommended. Some experts favor the use of epidural fibrin glue in an effort to avoid surgery for patients with a clearly identified site of cerebrospinal fluid (CSF) leak who have failed an adequate trial of repeated EBP. (See 'Epidural fibrin glue' above.)

•For patients with spontaneous intracranial hypotension who have failed an adequate trial of repeated EBP and in whom the site of the CSF leak cannot be identified, continuous epidural infusion of saline or dextran may be an option at centers with expertise in this technique. (See 'Continuous epidural infusion' above.)

●Recurrence – The estimated recurrence rate of spontaneous spinal CSF leakage is approximately 10 percent regardless of treatment. (See 'Prognosis' above.)