CEREBROSPINAL FLUID LEAKS: RHINORRHEA AND OTORRHEA
KEY POINTS
- Radionuclide cisternography is most useful to demonstrate that a leak is present but is only marginally useful for localization.
- The site of leakage can almost always be identified with multiplanar computed tomography and no intrathecal contrast.
- Computed tomographic cisternography is only necessary in unusually complex cases and in an active practice seeing such problems in <10% of cases.
- Magnetic resonance imaging is useful to identify or confirm associated or causative herniated brain and meninges.
Cerebrospinal fluid (CSF) rhinorrhea or otorrhea suggests that the subarachnoid space communicates with the nasal cavity or tympanomastoid cavities. Such a communication creates the potentially grave risk of intracranial infection; therefore, the condition must be diagnosed accurately and the communication sealed off. The diagnosis and localization a CSF leak is sometimes very challenging. There are also some common misunderstandings especially with regard to the utility and precise role of positive contrast computed tomographic cisternography. The interdisciplinary team does best when there is direct discussion of the specific task being asked of the diagnostic process. Proper communication and focused imaging will produce the best treatment options for the patients.
CSF leaks may be generally categorized as acquired or congenital. Acquired causes include trauma usually with fractures that result in a dural defect, postoperative leaks, infections, and benign or malignant neoplasms. Congenital CSF leaks are most often related to developmental bony pathways of communication and related dural deficiencies and more frank developmental abnormalities such as meningoencephaloceles (Chapter 79) and inner ear dysplasia (Chapter 106).
ANATOMIC AND DEVELOPMENTAL CONSIDERATIONS
Embryology
The developmental considerations that result in encephaloceles or lesser potential communications between the nasal cavity and skin and the subarachnoid space are discussed in Chapter 79 for the craniofacial region and in Chapter 106 for the temporal bone.
Applied Anatomy
The anatomy pertinent to CSF rhinorrhea is that of the cribriform plate and central skull base as it relates to the adjacent sinuses and nasal cavity; this is considered in Chapter 78.
The anatomy pertinent to CSF otorrhea is that of the inner ear, roof of the mastoid and middle ear, and temporal bone in general along with variants such as those related to arachnoid granulations that are discussed in Chapter 104.
IMAGING APPROACH
Techniques and Relevant Aspects
There really should be no attempted definitive anatomic imaging with computed tomography (CT) or magnetic resonance imaging (MRI) until there is reasonable clinical certainty that a CSF leak exists. This is easier to confirm when fluid is leaking onto the nasal cavity where fluid can be collected and tested than it is in the temporal bone where sampling entails entering the middle ear or mastoid in some manner. The presence of a leak at either site can also be confirmed by radionuclide cisternography (Chapter 5).
The general CT and magnetic resonance (MR) techniques used in this clinical setting are presented in Appendixes A and B. These searches require the highest possible multiplanar spatial resolution, and CT requires imaging processing at both bone and soft tissue algorithms. Radionuclide cisternography is discussed in Chapter 5.
Pros and Cons
Detecting a Leak
Screening begins clinically by testing for beta-2 transferrin in the nasal discharge or fluid recovered from the middle ear or mastoid. In the face of recurrent meningitis or other suspicion of leak, it is typical to confirm or exclude a CSF leak using radionuclide cisternography if fluid cannot be collected. This technique is also valuable in the setting of questionable symptoms of continued leakage if fluid cannot be collected after a repair. While the radionuclide cisternogram is a very sensitive imaging test for confirming CSF leakage as a cause for rhinorrhea or otorrhea, it is much less definitive for precise localization. The technique involves the injection of radiotracer into the thecal space with endoscopic pledget placement at the site of potential leak (Fig. 5.17). Positive contrast computed tomographic cisternography is really only able to confirm relatively high flow leaks; thus, its role in the detection of leaks is limited to problem cases such as those where there are multiple potential sources of a confirmed leak (Fig. 80.1).
Identifying of the Site of the Leak
Imaging plays a critical role in the management of CSF leaks by precisely identifying the site of the leak. The two most common areas for anatomic defects producing CSF rhinorrhea are the olfactory fossa along the vertical attachment of the middle turbinate and the superior and lateral walls of the sphenoid sinus. These defects can be detected by using careful multiplanar CT techniques without intrathecal contrast, at times combined with MR, about 90% to 95% of the time (Figs. 80.2–80.5). Intrathecal contrast computed tomographic cisternography may help find the site of the leak in very complex or problem cases, such as those with multiple potential sites of leakage or when the surgical approach may be complicated, and when the leak is active (Figs. 80.5–80.10). Sometimes, pooling contrast in the sinuses or nasal cavity or nasopharynx establishes that a leak is present, but most of the time in rhinorrhea this is redundant information since the presence of a leak has been established. In the temporal bone, where fluid may not be available, testing for a leak with positive contrast computed tomographic cisternography makes more sense.
MRI “cisternography” using highly T2-weighted or steady state images may be useful but is not preferred to CT as the initial study because it lacks the cortical bony detail necessary for a high accuracy in localizing CSF leaks (Fig. 80.3F). Intrathecal gadolinium can be used, but this is not a generally accepted option currently.
If a leak is detected, most surgeons will administer a dilute solution of intrathecal fluorescein and endoscopically examine the area to help identify the exact anatomic locations of the leak during the repair process. If the leak is not identified after the best attempts at imaging and there remains a strong clinical suspicion, intrathecal fluorescein may be utilized to detect and aid in the repair; this is usually combined with an endoscopic approach for repair of the skull base.
SPECIFIC DISEASE/CONDITION
Cerebrospinal Fluid Rhinorrhea
Etiology
CSF rhinorrhea results from a breakdown of the barriers that separate the subarachnoid space from the nasal cavity and/or paranasal sinuses. CSF rhinorrhea from the anterior cranial fossa is most common after head trauma followed by that due to endoscopic sinus surgery (ESS) (Figs. 80.1–80.4) or more extensive sinonasal operative procedures. The incidence of a CSF fistula after ESS is likely <1%. Conditions that increase the ventricular pressure may be important contributors to the pathogenesis of CSF leaks. Congenital causes include developmental meningoencephaloceles and might be extended to include arachnoid granulations or just developmental areas of larger than normal bony defects along the cribriform plate and central skull base that might lead to “spontaneous” CSF rhinorrhea (Figs. 80.1–80.3 A,B; 80.5; 80.9; and 80.10).
FIGURE 80.1. A patient with cerebrospinal fluid rhinorrhea. Computed tomographic cisternography was done. A: Coronal image showing contrast accumulating in the sphenoid sinus (arrowhead) with areas of dehiscence along the sphenoid roof not seriously considered a source of leakage. B: Coronal image at bone windows through the cribriform plate showing a subtle defect in the cribriform plate with intracranial structures somewhat lower on the right (arrow) than on the opposite side. A small stream of contrast (black arrowhead) may be visible. There also appears to be localized mucosal thickening in the upper nasal cavity on the right side and in the ethmoids on the right (black arrows) as possible evidence of localized irritation of the mucosa in this region due to leakage. C: An image slightly posterior to that seen in (B) showing any likely tract of contrast extending to or through the defect on the right side of the cribriform plate (arrow) and possible localized mucosal irritation (arrowhead). At surgery, the defect was identified and patched. (NOTE: Intrathecal contrast was used in this patient because there was more than one area suspicious as a site of leakage. The study also shows that accumulation of contrast in the sphenoid sinus cannot be taken as firm evidence of a leak within that sinus since fluid can accumulate, by gravity in the supine position, in the sphenoid sinus due to a leak from the cribriform plate.
Clinical Presentation
Spontaneous CSF leaks will present as an unusually watery, unilateral nasal discharge referred to as CSF rhinorrhea. It may also be discovered in an investigation for a cause of meningitis, especially recurrent meningitis caused by typical nasal-origin pathogens.
Pathophysiology and Patterns of Disease
Fractures and iatrogenic trauma produce defects in the anterior skull base that result in a dural tear and cause CSF leakage. Trauma may also cause a bony defect that predisposes the patient to developing a posttraumatic encephalocele or meningocele that in turn results in delayed CSF leakage (Fig. 80.3 C,D).
Some cases of spontaneous CSF rhinorrhea are caused by obvious skull base defects transmitting a developmental meningocele and/or encephalocele (Figs. 80.3D, 80.5, and 80.10). These may present in childhood or early adult life. Other more subtle congenital bony dural defects that result in CSF leakage occur along the cribriform plate, where small areas of dehiscence are normal (Fig. 80.1). Another common area of such natural dehiscence is in the sphenoid sinus when the sinus development results in prominent lateral extension of its air cells such that the course of the maxillary division of the trigeminal nerve is along the sinus roof. The canal/groove for the nerve then provides an area of bony weakness (Figs. 80.5 and 80.10). These natural areas of bony weakness are exposed to intermittent changes in pulsatile CSF pressure and the weight of the temporal lobe that likely cause further bony erosion over the course of many years. This results in CSF leaks most commonly associated with at least some herniation of brain and meninges (Figs. 80.5 and 80.10). Occasionally, there are patients who just seem to have deficient bone in the area that may predispose to multiple areas of such pathology. Surgeons may report that the dura in this area appears to be fenestrated almost to the point of appearing macerated.
FIGURE 80.2. This case illustrates one in which intrathecal contrast was ordered and used but was not necessary. A: Coronal section from a computed tomographic cisternography study showing accumulation of contrast below the cribriform plate and in the upper nasal cavity (arrows). B: Sagittal images clearly showing a relatively large defect in the cribriform plate (arrows) and contrast leaking into the area of this defect. This patient had an obvious high-volume leak clinically that based on just imaging was obviously through the cribriform plate; the use of intrathecal contrast was necessary.
As in the temporal bone, the central and anterior skull base may have developmental arachnoid granulations that contribute to this process. These arachnoid granulations may be aberrantly located over a pneumatized part of the skull rather than invaginated into the dural-covered peripheral intracranial venous drainage system, creating a bony weakness that when exposed to constantly pulsatile forces results in a defect that eventually produces a leak (Figs. 80.6 and 80.11). Any of these substrates for the development of a leak may be aggravated or accelerated by conditions that raise the intracranial CSF pressure.
Manifestations and Findings
Computed Tomography
In the acute trauma and postoperative settings, CT will show fractures and may show intracranial air as a clue to the presence of a related dural tear. It may show herniated brain and/or meninges. Fluid levels are rare.
In cases of spontaneous leaks, a polypoid mass suggesting a meningoencephalocele leading to an obvious skull base defect may be present (Fig. 80.3C). In more subtle cases, there may be an unusual morphology to the ethmoid roof and olfactory fossa region heralding subtle areas of abnormal dehiscence (Fig. 80.3A,B). There may be evidence of old fractures. In the central skull base, there may be an unusual configuration of the base of the sphenoid bone, floor of the middle cranial fossa, and/or prominent lateral recesses of the sphenoid sinus with thin or dehiscent bone usually very near the course of the maxillary division of the trigeminal nerve (Figs. 80.5–80.8).
Fluid levels are occasionally present. Fluid in the sphenoid sinus is much more commonly “collected” in the supine position from a nasal cavity leak than indicative of a sphenoid sinus leak unless prominent sphenoid sinus lateral recesses are present (Fig. 80.5A). It is more common to see localized mucosal thickening in the upper nasal cavity and adjacent sinus due to the irritating nature of leaking CSF (Fig. 80.1B). If the leak is high flow and intrathecal contrast is given, the contrast may be seen at the site of the fistula and/or collecting in the sinuses or related meningocele (Figs. 80.6–80.9).
Magnetic Resonance
MRI can easily confirm or exclude a meningoencephalocele (Figs. 80.3F, 80.4, 80.5, and 80.10). This is very useful in cases where there may be an unusual morphology to the ethmoid roof and olfactory fossa or along the central skull base sphenoid sinus region and a strictly endoscopic repair is being considered.
Nuclear Medicine
The use of radionuclide cisternography was discussed earlier in this chapter and in Chapter 5.
Endoscopy
It is typically very difficult to identify the site of a leak at diagnostic endoscopy; however, that examination may reveal a meningoencephalocele as a cause.
Differential Diagnosis
From Clinical Data
Spontaneous CSF rhinorrhea must be distinguished from other causes of rhinorrhea due to inflammatory or neurologic conditions of the nasal cavity, such as allergic or vasomotor rhinitis, respectively. Unilateral and very watery output raises the index of suspicion and should lead to screening by testing the fluid for beta-2 transferrin.
FIGURE 80.3. Three patients with cerebrospinal fluid leakage, none of which required intrathecal contrast for evaluation. A, B: Patient 1. In (A), the coronal section shows a dysmorphic appearance of the anterior skull base with a very deep olfactory fossa and bulging of meninges into this region (arrows) and prominent defects in the cribriform plate region, especially on the right, as the source of cerebrospinal fluid (CSF) leakage (arrowhead). In (B), bone windows at nearly the same position as (A) show the dysmorphic anterior skull base in association with a somewhat dysmorphic attachment of the superior turbinate (arrow) and possible irritation of the adjacent nasal septal mucosa due to CSF leakage (arrowheads). This area was successfully patched via endoscopic sinus surgery. C, D: Patient 2, who had a history of trauma. In (C), the coronal computed tomography study shows a polypoid mass projecting into the ethmoid complex through a large defect in the ethmoid roof (arrow). In (D), the coronal T2-weighted image shows the meningoencephalocele extending into the ethmoid sinuses (arrow). E: Patient 3 with a history of trauma and a disruption of the epitympanic roof (arrow) leading to CSF otorrhea. This leak resolved with conservative management.
Related posts:
Preseptal Compartment: Acute and Chronic Infections and Noninfectious Inflammatory Conditions
Necrotizing (“Malignant”) Otitis Externa
Hypopharynx: Developmental Abnormalities
Tumors of the Submandibular Gland and Space and Tumorlike Conditions
Masticator Space, Buccal Space, and Infratemporal Fossa Infections and Other Inflammatory Conditions
Oral Cavity and Floor of the Mouth: Malignant Tumors
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
