Misdiagnosis of Spontaneous Intracranial Hypotension FREE

A HEADACHE that worsens on assuming an upright position and is relieved by recumbency is a well-recognized complication of lumbar puncture.¹ When the onset of such a positional headache is spontaneous, a diagnostic challenge is created even though spontaneous intracranial hypotension has been recognized for more than 6 decades, and magnetic resonance imaging has greatly facilitated the diagnosis.^{2 - 12} A spontaneous spinal cerebrospinal fluid (CSF) leak is the typical cause of spontaneous intracranial hypotension.^{6 - 12} Such spontaneous CSF leaks may result from simple dural rents or fragile arachnoid cysts, often in the presence of an underlying but previously unrecognized generalized connective tissue disorder.^{7 ,13 - 16} The symptoms of spontaneous intracranial hypotension may become debilitating. Effective treatment is available, and the risks of treatment for the variety of conditions that may mimic spontaneous intracranial hypotension are significant. Therefore, the importance of an expeditious correct diagnosis in these patients cannot be overemphasized. I investigated the diagnostic delay in a cohort of patients with spontaneous intracranial hypotension.

The patient population consisted of a consecutive group of patients who were referred to the Maxine Dunitz Neurosurgical Institute (Los Angeles, Calif) for consideration of surgical repair of a spontaneous spinal CSF leak between January 1, 2001, and June 30, 2002. All patients were evaluated in an outpatient setting, a treatment plan was initiated, and clinical follow-up was obtained at regular intervals. Information regarding the initial clinical course and details of the various physician encounters was gathered by interviews with the patients and physicians by telephone or written correspondence and by review of medical records.

During the 18-month study period, 18 patients (15 women and 3 men; mean age, 38 years [range, 22-55 years]) were identified as having a spontaneous spinal CSF leak (Table 1). The CSF leak was documented by computed tomographic myelography in all patients. All patients had health insurance, and their level of educational attainment was high. Treatment consisted of surgical repair of the spontaneous spinal CSF leak in 13 patients and nonsurgical management, including epidural blood patching, in 5 patients.

All patients had a positional headache at initial examination. The headache was unlike any other they had ever experienced, and a variety of associated symptoms was present (Table 2). All but 2 patients remembered the exact time, date, and circumstances of the onset of the headache. Ten patients sought prompt medical attention (ie, within 24 hours from the onset of headache), and the remaining 8 patients waited 2 days to 2 months to seek medical attention (median, 3 days; mean, 22 days).

Only 1 patient was diagnosed as having spontaneous intracranial hypotension at the first physician encounter, and 17 patients (94%) initially received an incorrect diagnosis (Table 3). These 17 patients sought attention from 1 to 6 physicians before the correct diagnosis was made, and a total of 33 working diagnoses were established (Table 4 and Table 5). The diagnostic delay ranged from 4 days to 13 years (median, 5 weeks; mean, 13 months).

Diagnostic procedures associated with risk included cerebral arteriography in 2 patients, craniotomy with meningeal biopsy in 1 patient (Figure 1), and transesophageal echocardiography in 1 patient. Erroneous diagnoses resulted in craniotomies for decompression of cerebellar tonsillar descent (mimicking Chiari type I malformation) in 2 patients (Figure 2) and evacuation of subdural fluid collection in 1 patient (Figure 3).

In this study, the diagnosis of spontaneous intracranial hypotension was rarely made at the initial physician visit and typically patients endured weeks, months, or even years of treatment for a variety of neurologic disorders. Most likely, this was due to the unfamiliarity of physicians with the syndrome of spontaneous intracranial hypotension. Also, the findings from CSF examinations as well as the imaging features of spontaneous intracranial hypotension may mimic other, more serious neurologic conditions.

The typical patient with spontaneous intracranial hypotension has a headache that occurs or worsens soon after assuming an upright position and disappears or improves after resuming a recumbent position.^{2 - 16} The onset of the headache is usually gradual, reaching maximal intensity in several minutes to hours. However, some patients report an acute onset of headaches (thunderclap headache), and it is not surprising that such patients are suspected of having a subarachnoid hemorrhage and may undergo not only computed tomography but cerebral angiography as well.¹⁷ The severity of the headache varies, and many mild cases undoubtedly are never diagnosed. In some patients, however, the headaches become incapacitating, preventing them from engaging in any useful activity while in an upright position. The headache most likely is the direct result of the decrease in CSF volume and the downward displacement of the brain, causing traction on pain-sensitive structures. Although a postural headache is the clinical hallmark of intracranial hypotension, it is well known that the posture-related component often becomes less prominent or even disappears over time when the underlying CSF leak is left untreated and becomes chronic. It is an exceptional patient who has no posture-related component to the headache from the onset of symptoms.^{18 - 19}

A wide variety of symptoms may be associated with spontaneous intracranial hypotension, sometimes becoming more prominent or more worrisome than the headache, but the presence of the unique positional features of the headache should direct the physician to the correct diagnosis. Nausea, vomiting, photophobia, and posterior neck pain or stiffness are commonly seen and suggest meningeal irritation, mimicking subarachnoid hemorrhage or infectious meningitis. Many of the other associated symptoms are believed to be directly related to the downward displacement of the brain due to loss of CSF buoyancy. Diplopia, changes in hearing, and vertigo are common and may be caused by stretching of the ocular motor, cochlear, and vestibular nerves, respectively.^{2 - 18 ,20 - 23} Similar mechanisms could explain less commonly reported manifestations of intracranial hypotension, such as facial numbness or pain (trigeminal nerve), facial weakness or spasm (facial nerve), and dysgeusia (chorda tympani).^{7 - 8} Transient visual obscurations, visual blurring, or visual field defects also have been reported and may be attributed to the stretching of the optic apparatus over the pituitary fossa.²⁰ Stretching of cervical nerve roots could be implicated as a cause of the radicular arm pain or arm numbness occasionally found in patients with spontaneous intracranial hypotension. An alternative hypothesis to explain the disturbances of hearing or balance is that the abnormal change in CSF pressure is directly transmitted to that in the cochlea or labyrinth.²⁴

Rarely, severe sagging of the brain may result in coma from diencephalic or hind brain herniation.^{25 - 27} Other rare manifestations may include parkinsonism and dementia.^{28 - 29} More subtle cognitive deficits, however, are probably underrecognized and may not be appreciated until cognition improves following successful treatment of the spinal CSF leak.

Typically, patients with spontaneous intracranial hypotension have an opening pressure at the time of lumbar puncture of less than 60 mm of water (normal, 65-195 mm of water).⁶ Occasionally, a dry tap is initially encountered, and CSF can only be obtained with aspiration using a syringe, with the patient in an upright position, or with a Valsalva maneuver. Because of the difficulties in obtaining a CSF sample, a traumatic tap is not infrequent, contributing to the confusion with regard to a diagnosis of subarachnoid hemorrhage. Examination of CSF often demonstrates abnormal findings, such as an elevated protein content and pleocytosis.^{6 - 10} The pleocytosis is primarily lymphocytic, and the white blood cell count may exceed 200 white blood cells per milliliter, a finding that may be confused with infectious meningitis.^{6 - 10}

Magnetic resonance imaging of the head has now supplanted lumbar puncture as the diagnostic study of choice in the initial evaluation of patients suspected of having spontaneous intracranial hypotension. The 3 most characteristic imaging features of intracranial hypotension are enhancement of the pachymeninges, downward displacement of the brain (brain sagging), and subdural fluid collections.^{6 - 13 ,30}

Most, but not all, patients with spontaneous intracranial hypotension demonstrate pachymeningeal enhancement on magnetic resonance imaging.^{6 - 13 ,30 - 31} The enhancement is diffuse, only involves the pachymeninges (dura mater), and spares the leptomeninges.^{6 - 13 ,30 - 31} There may be diagnostic confusion with neoplastic, inflammatory, or granulomatous disease. It is not uncommon for patients with spontaneous intracranial hypotension to have undergone a meningeal biopsy. Reactive hyperemia of the pituitary gland may also be seen, mimicking a pituitary tumor,^{32 - 33} and may occasionally be associated with a concurrent increase in the level of circulating prolactin.³⁴

Brain sagging is a very specific finding in spontaneous intracranial hypotension.^{7 - 10 ,30} It can be identified by several features, such as effacement of the suprasellar cistern, bowing of the optic chiasm over the pituitary fossa, flattening of the pons against the clivus, and downward displacement of the cerebellar tonsils. The displacement of the cerebellar tonsils may be mistaken for a Chiari I malformation, and some patients have undergone decompressive posterior fossa surgery, with no relief of their symptoms.⁷

Subdural fluid collections are common in patients with spontaneous intracranial hypotension.^{7 - 12 ,18 - 19 ,35} Most of these fluid collections are bilateral, thin, and do not cause any appreciable mass effect. Occasionally, larger symptomatic, subdural hematomas require surgical evacuation, but if the underlying spinal CSF leak is left untreated, the recurrence rate of the subdural hematoma is high.³⁵ Loss of CSF volume may also result in compensatory enlargement of the spinal epidural venous plexus, which may become symptomatic and mimic neoplastic, inflammatory, or degenerative disc disease.^{36 - 39}

Most patients with spontaneous intracranial hypotension respond well to bed rest, a generous intake of oral fluids, and the administration of caffeine, glucocorticoid medication, or mineralocorticoid agents. The mainstay of treatment of persistent spontaneous intracranial hypotension is the placement of 1 or more lumbar epidural blood patches, using approximately 10 to 15 mL of blood. If the patient fails to respond to this therapy, computed tomographic myelography is performed to confirm and localize the CSF leak. In most patients, a CSF leak is found in the cervical or thoracic spine. Therefore, consideration should be given to the placement of a high-volume (20-40 mL) thoracolumbar epidural blood patch, after which the patient is placed supine in the Trendelenburg position for 20 to 30 minutes, followed by the prone position for a similar time period. This allows blood to travel over many spinal segments toward the site of the CSF leak. If this fails to alleviate the symptoms, a directed epidural blood patch or percutaneous placement of fibrin glue is recommended. At the Maxine Dunitz Neurosurgical Institute, surgical treatment is reserved for those patients in whom these nonsurgical measures have failed. Computed tomographic myelography is the diagnostic test of choice to precisely localize the CSF leak,⁷ but a retrospinal fluid collection at the C1-C2 level should not be mistaken for the actual site of CSF leakage.³⁶ Rather, these C1-C2 CSF collections are the result of CSF leaking into the epidural space at another site, extending rostrally within the spinal canal and eventually escaping into the soft tissues at the C1-C2 level.³⁷

These data clearly show that for most patients with spontaneous intracranial hypotension, the diagnosis is missed initially and the diagnostic delay is significant. This exposes patients to risks associated with treatment for the disorders that can mimic intracranial hypotension. It is likely that for other patients with spontaneous intracranial hypotension, the diagnostic delay is even more pronounced, or the correct diagnosis is never established. It should be noted that the patients in this report were well educated, had medical insurance with ready access to medical care, and had symptoms severe enough to warrant consideration of surgical correction of the underlying spinal CSF leak and thus were more likely to be diagnosed. Headache is one of the most common reasons for a consultation with a primary care physician, a visit to the emergency department, or referral to a neurologist.⁴⁰ Clearly, spontaneous intracranial hypotension is rare among all patients with headache complaints, but it is an important diagnostic consideration for patients with "new daily persistent headaches."^{41 - 42} Increasing the awareness of this spontaneous form of positional headaches will likely result in fewer patients in whom this diagnosis is missed.

Corresponding author and reprints: Wouter I. Schievink, MD, Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical Center, 8631 W Third St, Suite 800 E, Los Angeles, CA 90048 (e-mail: schievinkw@cshs.org).

Accepted for publication July 28, 2003.