Anatomy of the anterior skull base

Surgeons must understand the relationship between the ethmoid sinuses and the anterior skull base. Understanding this relationship begins in the preoperative setting through reviewing CT images of the patient’s anatomy. CT imaging must be performed in the coronal plane in all patients before undergoing endoscopic sinus surgery. If available, thin slice (1.0–3.0 mm) axial images reconstructed in the coronal and sagittal planes provided further elucidation of the skull base anatomy.

Coronal images provide crucial information on several critical areas that, if overlooked, can lead to intraoperative complications. On coronal images, the height of the maxillary sinus should be compared with the height of the ethmoid sinuses. The ratio of maxillary-to-ethmoid height ( Fig. 1 ) can vary from 1:1 to 2:1 and can be used preoperatively to guide how dissection through the basal lamella will lead to the anterior skull base along the posterior ethmoid sinuses. Meyers has shown how this ratio can lead to inadvertent injury to the anterior skull base during dissection. The ratio can be further examined along the slope of the skull base, which is best viewed in sagittal or parasagittal images. A line extended from the limen nasi to the basal lamella will identify the distance to which the skull base will be encountered. Additionally, the slope of the skull base from frontal recess to planum sphenoidale will correlate these findings. The thickness of the anterior skull base should also be examined thoroughly with both coronal and sagittal CT images. The thickness of the anterior skull base varies and is often asymmetric in individual patients. Knowing the thickness of the anterior skull base is important, because the roof of the ethmoid has been shown to be the place of least resistance during dissection ( Fig. 2 ) . In addition to its anatomy, the skull base itself should be examined for areas of dehiscence, thinness, and thickness, and for soft tissue that might suggest an encephalocele.

A coronal CT of a preoperative patient demonstrating a maxillary sinus-to-ethmoid sinus ratio of 2:1. The short ethmoid height serves as an indicator to the surgeon that the anterior skull base will be quickly encountered when the basal lamella is removed before dissection of the posterior ethmoid sinuses.

A sagittal CT of the maxillofacial region in a patient after endoscopic sinus surgery in which the anterior skull base was injured during posterior ethmoid dissection, the point where the skull base is the weakest. This patient had a maxillary sinus-to-ethmoid sinus ratio greater than 1:1. The figure illustrates the “angle of attack” that is directed through the basal lamella.

The region of the ethmoid roof and cribriform plate is of critical importance and is best analyzed in the coronal plane. In the midline of the ethmoid bone is the cribriform plate ( lamina cribrosa ), which represents the lowest point of the skull base in the median aspect of the nasal cavity. Multiple perforations of the cribriform plate permit transmission of neural fibers into the nasal cavity that contribute to olfaction. The lateral mass (or labyrinth) of the ethmoid is suspended laterally from the cribriform plate by the fovea ethmoidalis, or ethmoid roof. Interposed between the cribriform plate and the fovea ethmoidalis is the lateral cribriform lamella. The sagittal buttress of the middle turbinate is suspended from the articulation of the cribriform plate and the lateral cribriform lamella. When viewed in the coronal plane, the three bony structures are said to possess a “gull wing in flight” appearance ( Fig. 3 ); however, this structure has significant variations. Keros and colleagues studied the relationship of these bony structures and proposed a classification based on the length of the lateral cribriform lamella. The classification is as follows: Keros type I (1–3 mm deep), Keros type II (4–7 mm deep), and Keros type III (8–16 mm deep). Ali and colleagues showed that type II anatomy is the most commonly encountered and the symmetry of the olfactory fossa varies widely among patients . The classification is useful because patients who have a lower Keros classification are believed to have a lower risk for intracranial injury and those who have a higher Keros classification a higher risk for intracranial injury and postoperative CSF leak.

A coronal CT image of a patient showing a low and asymmetric olfactory fossa. Dissection along the right anterior skull base can lead to inadvertent injury if the surgeon does not recognize the differences in ethmoid height in this patient.

Intraoperative management of cerebrospinal fluid leaks

Despite review of the preoperative films, use of intraoperative stereotactic image-guided surgery, and meticulous surgical technique, a violation of the anterior skull base can occur. If this untoward event occurs, the authors recommend the algorithm provided in Fig. 4 to manage this complication.

Algorithm for addressing intraoperative CSF leaks.

When a leak occurs, the surgeon should stop and review the anatomy. Landmarks (eg, posterior wall of the maxillary sinus, attachments of the middle turbinate, anterior face of the sphenoid sinus, skull base) within the patient should be identified to help reorient the surgeon, and then the preoperative and stereotactic imaging should be used to help localize the probable site of the leak. The most likely sites of injury include the lateral cribriform lamella and cribriform plate; therefore, surgeons should review these areas for dehiscences, asymmetries, thinness/thickness, or any bony abnormalities on the preoperative imaging. Additional attention should be directed to the posterior ethmoid sinuses to determine if reduced posterior ethmoid height (ie, maxillary–ethmoid ratio greater than 1:1) is present that would precipitate inadvertent injury to the skull base. If a particular region is suspected, an attempt should be made to correlate the radiographic site with the region endoscopically.

If the site cannot be completely identified or if the operating surgeon does not feel equipped to repair the defect, then adequate hemostasis should be obtained and the nasal cavity lightly packed to help apply gentle pressure to the defect. The surgeon should be aware that the process of gaining hemostasis can actually cause more complications, and the endoscopic view greatly magnifies any blood loss. They should therefore proceed with caution, especially when using monopolar cautery near the skull base. The upper and lower aerodigestive tract should be aspirated for blood and antiemetics administered to minimize postoperative nausea. All efforts should be made to minimize increases in intracranial pressure, and particular attention should be directed toward minimizing Valsalva maneuvers during arousal from general anesthesia. If safe, a deep extubation can be performed. Avoidance of nasal positive airway pressure is imperative, despite the extubation plan, because of the potential for pneumocephalus. Immediate head CT is warranted to document that no hematoma or pneumocephalus has occurred, which can be life-threatening.

The efficacy of antibiotic prophylaxis for preventing meningitis in patients who have postsurgical CSF leaks has not been studied in a prospective manner; however, several retrospective reviews of traumatic CSF leaks have yielded conflicting information. A meta-analysis by Brodie and colleagues showed no significant differences in the rates of ascending meningitis in patients treated with and without antibiotics in traumatic anterior skull base CSF leaks. Conversely, Bernal-Sprekelsen and colleagues found an incidence of ascending meningitis in 29% of patients treated conservatively (eg, head of bed elevation, bed rest). Although not statistically significant, 40% of patients treated transcranially also developed meningitis. However, in a separate study, these investigators found an incidence of ascending meningitis in 36.5% of patients undergoing endoscopic repair of CSF leaks. During endoscopic sinus surgery, presumably the anterior skull base is penetrated with an instrument (eg, curette, microdebrider), and bacteria or inflamed or infected mucosa is implanted intracranially, which may contribute to infection. Therefore, the authors believe it is prudent to administer perioperative parenteral antibiotics (eg, ceftriaxone) that effectively cross the blood–brain barrier, because an intraoperative CSF leak by definition includes an element of direct trauma to the brain, however minor.

A subarachnoid lumbar drain has been advocated for planned treatment of CSF leaks. However, no prospective studies have examined the use and duration of lumbar drains in patients who have experienced iatrogenic endoscopic sinus surgery injuries. In most instances, the authors believe a lumbar drain is unnecessary for an iatrogenic injury unless the patient has signs of benign intracranial hypertension, empty sella syndrome, or radiographic evidence of dehiscent bone along the skull base that would otherwise suggest elevated intracranial pressures.

Finally, transfer of service to a local rhinologist is recommended for continued care of the patient. Operating surgeons should proceed with closure only after they have identified the site of leak, understand the anatomy, and are prepared to attempt closure, and the patients are stable.

Transnasal endoscopic repair of CSF leaks has been reported to be successful in 90% to 97% of patients during a first attempt. Once the site of the leak is identified and prepared, surgeons should adhere to three goals: (1) safe and successful closure of the leak; (2) maintenance of sinus function; and (3) prevention of postoperative complications. The major sites of anterior skull base injury are reviewed.

Ethmoid/cribriform leaks

The surgical site is prepared by completing a maxillary antrostomy and total ethmoidectomy with wide exposure of the skull base. If exposure of the defect requires the removal of the middle turbinate, then its resection should be complete to prevent postoperative lateralization of the middle turbinate remnant and iatrogenic frontal recess obstruction. If the middle turbinate is removed, it should be preserved for later use as grafting material. If the repair is likely to cause postoperative obstruction of the frontal recess or the sphenoethmoidal recess, a prophylactic frontal sinusotomy or sphenoidotomy should be performed.

Once the defect is fully delineated, the mucosa surrounding the defect is stripped and potentially fulgurated with bipolar electrocautery 3 to 4 mm beyond the bony defect rim. The authors then size the defect measured using a trimmed ruler. Defects smaller than 2 mm will likely heal with a combination of osteoneogenesis and soft tissue fibrosis; therefore, the authors do not recommend preparing the epidural space or placing a bone graft to avoid increasing the size of the defect. Simple abrasion of the surrounding bone (scraping versus drilling with diamond bur) with placement of an overlay graft (eg, mucosa, temporalis fascia) dressed with a thin layer of fibrin glue sufficiently patches most leaks. Excessive fibrin glue should be avoided to help facilitate remucosalization.

A composite graft is often suited for the repair of defects larger than 2 mm but smaller than 6 mm. In these circumstances, a middle turbinate composite (mucosa/bone) graft is an excellent choice for repair, because the bone of the middle turbinate has similar density to the bone of the cribriform plate and anterior skull base. Resection of the middle turbinate should occur at the skull base to prevent postoperative lateralization of the middle turbinate remnant and iatrogenic stenosis of the frontal recess. Once the turbinate is harvested, it can be filleted and the mucosa stripped from one side. Once the defect is measured, the excess bone of the middle turbinate is removed, taking care to leave behind mucoperiosteum. The appropriately sized middle turbinate bone is then wedged into the defect, allowing the periosteum to be applied to the cleaned skull base and the mucosa facing the nasal cavity.

Defects larger than 6 mm are closed in a multilayered fashion with underlay bone grafting to reconstitute the native skull base anatomy ( Fig. 5 ). Septal bone and mastoid cortex are two choices for bone grafting. The advantages of septal bone include its ease of access intranasally and the osteoneogenic properties of the overlying mucoperichondrium. Care must be taken when harvesting the septal bone, because iatrogenic septal perforations may result; furthermore, portions of the septal bone may be fairly flimsy and fracture easily during placement. In situations of poor septal bone, the authors recommend using mastoid cortex, because sufficient bone is available for contouring, and soft tissue (eg, fascia, muscle) can be obtained for additional layered support. Harvesting mastoid cortex, however, requires a separate incision and site preparation, which may increase intraoperative time. The final defect size should be measured and the mucosa surrounding the defect stripped and potentially fulgurated. The dura is elevated using a round knife, ball-tipped seeker, or other appropriate instrument, and the bone graft is placed in an underlay fashion. Fibrin glue is applied lightly, and a pliable soft tissue graft (eg, fascia, mucosa) is placed as an overlay and secured with a thin coating of fibrin glue.

Endoscopic view of an anterior skull base defect that is repaired with mastoid cortex bone as an underlay graft ( A ) and with septal mucosa as an overlay ( B ) graft.