Complications in Endoscopic Sinus Surgery
Despite numerous advances in surgical technique and instrumentation, the risk of serious complication during endoscopic sinus surgery (ESS) is ever present. This is an unavoidable result of the close proximity of various critical structures, such as the orbit, internal carotid arteries, and skull base, to the paranasal sinuses. It is the responsibility of the operating surgeon to minimize the risks through meticulous preoperative preparation, careful operative technique, and diligent postoperative care. Complications in ESS may be subclassified in several ways, including classifications based on anatomical location, severity, and timing. The first two methods are generally considered to be the most relevant in ESS.
With regards to anatomical location, adverse events are categorized according to the site or tissue involved; these include vascular, neurologic, ophthalmic, intranasal wound healing, facial, and packing-related complications. In the severity-based approach, events are broadly divided into major and minor complications. Major ESS complications are those that result in or carry a significant risk of long-lasting or permanent sequelae. Major complications that involve the cranial vault include cerebrospinal fluid (CSF) leak, tension pneumocephalus, meningitis, abscess, intracranial hemorrhage, direct brain injury, and encephalocele formation. Similarly, major complications may involve the eye, such as in medial rectus injury, optic nerve injury, orbital hematoma, and nasolacrimal duct injury, and may result in double vision, loss of vision, and epiphora. Complications that arise from damage to regional blood vessels—including the anterior or posterior ethmoidal, sphenopalatine, or internal carotid arteries—are considered major if the resulting hemorrhage affects cerebral circulation or causes a significant drop in hemoglobin level or requires a transfusion of red blood cells. Other major complications worth mentioning are anosmia and toxic shock syndrome, and complications of ESS may be fatal. Fortunately, all of these complications are quite rare. However, these risks should be disclosed in a thorough preoperative consent procedure due to the serious repercussions that may result.
Minor ESS complications are far more common. They may require surgical correction but generally do not produce any significant long-term adverse outcomes. These may include periorbital emphysema, ecchymosis, and fat herniation through a defect created in the lamina papyracea. A small amount of perioperative bleeding, not requiring transfusion, may be considered minor, as are facial swelling, hyposmia, hypesthesia of the infraorbital nerve or teeth, synechia formation, myospherulosis, atrophic rhinitis, and osteitis.
Finally, complications may be designated as intraoperative, early postoperative, or late postoperative. One example of an intraoperative complication is CSF leak. As with most other intraoperative complications, it is best if the surgeon recognizes the CSF leak at the time of the initial surgery so that an immediate repair can be performed and the risk of serious sequelae such as pneumocephalus, meningitis, or intracranial abscess are minimized. Early postoperative complications may include infection, hemorrhage, or adhesion formation; these may occur anywhere from immediately after surgery to 2 weeks following the procedure. Late complications may include mucocele or mucopyocele formation, and may occur many years after the procedure. It should be noted that all of the above categorizations are artificial and subject to interpretation. They are meant to be used with a degree of common sense and to aid in communication with patients and other clinicians.
Several risk factors exist for the occurrence of complications in ESS;1 these can be broadly divided into anesthetic, surgeon-related, and disease-related factors. General anesthesia increases the risk of complications through the lack of patient feedback when approaching sensitive structures like the lamina papyracea and the skull base. Also, certain forms of general anesthesia—inhalational in particular—tend to have an adverse effect on intraoperative bleeding and therefore visualization. Surgery on the right side of the nose is a risk factor for a right-handed surgeon because of the angle of the endoscope and instruments, as is left-sided surgery for a left-handed surgeon. Lack of surgeon experience is a risk for the obvious reasons of unfamiliarity with the anatomy and use of instrumentation. Finally, extensive sinus disease, excessive bleeding, and revision surgery, all of which can obscure or distort the sinonasal structures normally encountered during surgery, are important risk factors.
The current chapter seeks to provide a comprehensive and organized discussion of the complications that can occur during or following ESS. Information pertaining to the definition and incidence of the complication will be provided whenever possible. The chapter is subdivided based on the type of complication, and the precise moment during surgery when the complication is at risk of happening will be explained. The management of each complication will be discussed based on the best available evidence, and finally, suggestions for avoiding the complications will also be presented.
Special Concerns for Each Sinus
The Maxillary Sinus
The maxillary sinus is the most commonly addressed sinus in ESS. The uncinectomy, with or without an additional antrostomy, is often the first step in the endoscopic sinus procedure. Whether a limited or an extensive procedure is necessary, the maxillary sinus provides a reliable landmark for the initiation of and progression through the steps of ESS. Furthermore, controlling disease within the maxillary sinus is essential to achieving a successful surgical outcome. Although not generally considered a challenging area to operate on, an in-depth understanding of possible complications involving the maxillary sinus and their management is necessary for the above reasons.
Anatomical relationships of importance to maxillary sinus surgery include the orbit superiorly, the nasolacrimal duct anterior to the maxillary ostium, the infraorbital nerve running in or below the roof of the sinus, and the sphenopalatine foramen and artery posterior to the fontanelle of the sinus ( Fig. 9.1 ).
The Frontal Sinus
The frontal sinus has long been considered one of the most challenging areas to access and manage surgically. With significant patient-to-patient variability in pneumatization, anatomical complexity, and disease characteristics, a wide repertoire of surgical techniques is necessary for the sinus surgeon dealing with frontal sinus disease.
The frontal sinus is bounded anteriorly by the thick anterior table of the frontal bone, with its overlying periosteum, facial muscles, galea aponeurotica, soft tissues, and skin; and posteriorly by the much thinner posterior table, and dura of the anterior cranial fossa. Inferiorly are the orbital plates of the frontal bone laterally and the frontal beak medially; and superiorly is the frontal bone as it forms the anterior calvarium. The frontal recess is an hourglass-shaped space connecting the frontal sinus to the ethmoid infundibulum ( Fig. 9.2 ). Its boundaries are the skull base posterosuperiorly, the roof and posterior wall of the agger nasi cell anteroinferiorly, the orbit laterally and the middle turbinate medially. The frontal sinus drainage pathway passes through this area and owes its configuration to the fronto-ethmoidal cells that pneumatize into it. A precise understanding of each patient′s anatomical variations is essential for the safety and long-term success of their frontal recess dissections.
The Ethmoid Sinuses
The ethmoid sinuses are the most anatomically complex and variable of the paranasal sinuses; nevertheless, many ethmoidal air cells can be identified on preoperative imaging and intraoperatively in a very predictable manner. Adequate removal of the obstructing bony partitions and disease material from within these sinuses is key to the control of chronic rhinosinusitis in medically refractory patients. However, given their location, bordering on the orbit laterally and comprising a significant portion of the anterior cranial fossa floor superiorly, the potential for disastrous complications is particularly present when performing ESS in the region of the ethmoids ( Fig. 9.3 ). Furthermore, the majority of iatrogenic CSF leaks are created in the ethmoid region. Preoperative appreciation of high-risk anatomical variants is of particular importance for avoiding complications when performing surgery within the ethmoid region.
The Sphenoid Sinus
The sphenoid sinus is the most posterior of the sinuses, and is the gateway to the sella turcica in endoscopic transsphenoidal approaches to pituitary tumors. Wide sphenoidotomy also allows the operating surgeon to identify with confidence the level of the skull base in ESS procedures, and so it is often performed in more severe cases of chronic rhinosinusitis. Its close relationship to numerous critical structures, such as the brain, the optic nerves, the carotid arteries, the cavernous sinuses and cranial nerves contained therein ( Fig. 9.4 ), make it more prone to severe infectious and inflammatory complications in sinonasal disease. For this reason, it is recommended to treat severe sphenoidal disease more aggressively with surgical evacuation and debridement. However, it is also this close anatomical relationship to critical structures that predisposes it to disastrous operative complications, and correspondingly causes many surgeons to shy away from sphenoid surgery.
Excessive bleeding as a result of ESS may occur during the procedure or in the postoperative period; the majority of bleeding events occur early in the postoperative course. Excessive intraoperative bleeding may significantly affect the operator′s visual field, obscuring landmarks and predisposing to further operative complications. At the very least, bleeding may slow the progress of surgery, but in severe cases, it may force the surgeon to abort the procedure. The reported incidence for epistaxis requiring intervention are 0.6 to 1.6%,2 and for major hemorrhage requiring transfusion is 0.76%.3 Causes of bleeding can be surgical/technical or patient related. The bleeding may be diffuse or localized, and if severe enough it may significantly affect the operative visual field and render surgery unsafe.
Any hemorrhage that is severe enough to require a blood transfusion is considered a major complication of surgery.
Patient factors causing increased surgical bleeding can be broadly divided into local processes such as infection, or systemic comorbidities. The latter must be sought and addressed during the preoperative evaluation. Individuals with underlying hypertension, or peripheral vascular disease should be optimized preoperatively, as controlled hypotension may be more challenging in this group of patients. Liver and renal diseases are important causes of clotting factor deficiency and platelet dysfunction, respectively. Chronic alcohol abuse, malnutrition, and vitamin deficiencies (most notably vitamin K), can also affect coagulation, and must be explored if suspected clinically. Bleeding diatheses, such as hemophilia A or B, and von Willebrand disease, require clotting factor replacement or specialized pharmacotherapy (desmopressin, tranexemic acid, or aminocaproic acid), and must be planned for. Inherited collagen and blood vessel abnormalities, including hereditary hemorrhagic telangiectasias and arteriovenous malformations, are other less common causes of profuse surgical bleeding. Medications that will increase surgical bleeding must be halted before surgery; these include nonsteroidal anti-inflammatory drugs, aspirin, coumadin, heparin, and anti-platelet agents. Several herbal and alternative therapies can affect coagulation pathways, such as ginseng, ginkgo, kava, and fish oil. It is prudent for patients to discontinue all herbal and alternative medicines at least 7 days before surgery.
Important measures for optimizing the visual field in ESS include placing the patient in reverse Trendelenburg positioning, and the application of topical and local vasoconstrictors.4,5 Furthermore, studies have demonstrated that the ideal mean arterial pressure has been shown to be less than 60 to 75 mmHg and the ideal heart rate to be less than 60 beats/min.6,7 This is best achieved through the use of total intravenous anesthesia.6,8 Systemic steroids are helpful in the preoperative period to reduce the size and vascularity of polyps in patients with significant nasal polyposis, so reducing capillary bleeding during functional ESS.9 The use of perioperative antibiotics in patients with acutely infected and inflamed sinuses may also be beneficial. However, further studies are necessary to clarify the optimal dose, optimal length of treatment, and groups of patients that would benefit from these treatments.
At the end of every ESS procedure, the anesthetist should be asked to restore the patient′s blood pressure toward their preoperative level while hemostasis is verified. Thorough saline irrigation and suctioning of blood from within the nose and sinus cavities is performed, followed by careful inspection of the nasopharynx looking for pooling of fresh blood. A suction bipolar instrument is important, as it will allow simultaneous evacuation of blood from the operative field and cauterization of the bleeding points. Most of the mucosal oozing will stop spontaneously after sinus surgery; however, larger blood vessels can result in significant postoperative hemorrhage, especially if these vessels are in spasm during surgery and not dealt with at that time. Common areas of arterial bleeding after ESS that should be attentively examined are the regions supplied by large branches of the sphenopalatine artery and anterior ethmoidal artery. These regions include the posterior rim of an enlarged maxillary antrostomy, the area of the sphenopalatine foramen especially if a middle turbinate resection has been performed; additionally, the anterior face of the sphenoid sinus (which is supplied by the posterior nasal/septal branch) and anterior skull base should be carefully inspected. After appropriately cauterizing the points of bleeding, the nasopharynx is again suctioned and inspected for pooling of fresh blood; this should be the last maneuver performed in any endoscopic sinus procedure.
Nasal packing is not usually necessary after sinus surgery with proper hemostasis; however, some surgeons still employ this practice. Some alternatives to petroleum gauze packing include bioresorbable materials such as Nasopore (Stryker Corp., Kalamazoo, MI, USA), injectable gels, or finger cot dressing (sterile glove fingers stuffed with rolled 5 × 5-cm gauze sponges). These can either be removed in the immediate postoperative period, or left in the patient for later removal in clinic; the length of time they are left is dependent on the material used, with bioresorbables able to stay the longest.
More severe postoperative hemorrhaging may require aggressive management, beginning with attention to the patient′s “ABCs.” It is exceedingly rare that the patient′s airway and breathing require any intervention or protection due to heavy bleeding, but if this is judged to be the case, intubation, sedation and ventilation can be performed. Nasal packing in the postoperative setting is usually traumatic to the fragile healing tissues, and should be avoided when possible. A trial use of an inflatable hemostatic device, such as the Rapid Rhino (ArthroCare ENT, Austin, TX, USA), can be attempted as a temporizing measure until more definite management is undertaken. Depending on the region that is bleeding, ligation of the sphenopalatine or anterior ethmoid arteries may need to be performed. If the former is to be performed, greater palatine foramen infiltration with local anesthetic and epinephrine is helpful.5 The foramen can be palpated along the posterior hard palate, halfway between the palatal midline and the second molar tooth. A cadaver study performed by the senior author has shown that the successful injection is best achieved by bending a 25-gauge needle through 45° ~ 2.5 cm from the tip of the needle.10
Anterior Ethmoidal Artery Injury
The anterior ethmoidal artery (AEA) is a terminal branch of the internal carotid artery, which branches off the ophthalmic artery within the orbit and then heads medially to traverse the lamina papyracea and enter the anterior ethmoid sinus. The AEA usually runs in the skull base along the roof of the ethmoid, just posterior to the anterior face of the bulla ethmoidalis, giving off branches that feed the superolateral nasal wall. It should be visible after thorough clearance of the skull base in this region ( Fig. 9.5 ). It then enters the anterior cranial fossa through the lateral wall of the olfactory recess. From there, the artery traverses the cribriform plate to supply the anterior septum. Cadaver studies have shown that the AEA runs in a bony mesentery, hanging below the skull base, in approximately one-third of cases and is amenable to surgical clipping in ~ 20% of cases.11 Using endoscopic bipolar instruments, however, the artery may be adequately cauterized in a significantly higher percentage of patients.
Bleeding from the anterior artery may cause a significant hemorrhage during surgery, but more importantly, complete transection of the artery may result in retraction of the proximal (lateral) severed end into the orbit. The consequence of this is a rapidly expanding orbital hematoma, which is a surgical emergency; it is covered in more detail below.
The position of the AEA is identified on preoperative imaging before every surgical procedure, to determine whether it is in a bony mesentery or running within the ethmoid roof itself. The artery can be found on a coronal computed tomography (CT) scan as a pinch or “nipple” between the medial rectus and superior oblique muscles ( Fig. 9.6 ). Those running in a mesentery tend to be associated with a longer lateral lamella of the olfactory fossa (Keros, type 2 or 3) and a high ethmoid skull base. Care must be taken not to sever the AEA when using the microdebrider to clear polypoid tissue and bone from the ethmoid roof. An important technical point to prevent transection of the artery is to avoid passing the microdebrider blade in a posterior to anterior direction with the tip near the skull base ( Fig. 9.7a ); a safer alternative is to advance and withdraw the tip at an angle tangential to the skull base until the excess tissue and bone have been removed ( Fig. 9.7b ). This will result in a gentle debridement and will avoid the occurrence of a through and through cut of the artery. If necessary, a partial injury to the artery can be easily managed with suction bipolar diathermy. A complete transection of the AEA with retraction of the proximal bleeding end into the orbit should immediately raise concerns for rapid orbital hematoma formation; the management algorithm is presented in Figure 9.22 in the Orbital hematoma section below.
Sphenopalatine Artery Branch Injury
The sphenopalatine artery supplies a significant portion of the nasal cavity, and is often a source of arterial bleeding during ESS. The branch most often encountered is the posterior nasal artery as it branches off the sphenopalatine artery and runs along the anterior wall of the sphenoid sinus below the sphenoid ostium to eventually become the posterior septal artery ( Fig. 9.8 ). It is usually transected if the anterior sphenoid face is removed in a superior to inferior direction starting from the natural ostium with a downward biting instrument. If a wide sphenoidotomy is necessary for severe disease, the use of a sharp Kerrison or Hajek–Koeffler punch leads to a clean transection of the vessel, favoring hemostasis through effective spasm of the vessel after injury. In this way, brisk arterial bleeding may only be seen very briefly as the vessel rapidly goes into spasm; however, the artery may open up again in the early postoperative period and lead to significant hemorrhage. It is therefore recommended that the proximal and distal ends of the transected posterior nasal artery be electrocauterized at the end of surgery whenever a large sphenoidotomy is performed. If the posterior nasal mucosa is excessively traumatized during surgery and the artery is avulsed or partially cut, then arterial spasm will be less effective at achieving hemostasis and immediate cauterization of the bleeding at the time of injury will be necessary.
Because of the proximity of the sphenopalatine artery main trunk and the presence of feeding branches such as the posterior lateral nasal artery and the inferior turbinate branch,12 the area of the posterior fontanelle of the maxillary sinus is at risk of significant bleeding during enlargement of antrostomy. Additionally, vessels branch off the sphenopalatine artery as it emerges from the sphenopalatine foramen to supply the inferior and middle turbinates. These vessels are largest at the posterior end of the turbinates and run in the mucosa medial to the bone of the corresponding turbinate ( Fig. 9.8 ); it is therefore common to encounter brisk bleeding when performing turbinate reduction or turbinectomy involving these areas. Bipolar diathermy of the posterior ends of the turbinates is recommended whenever damage of these vessels is suspected, even if not actively bleeding at the end of the procedure, to prevent a latent hemorrhage.
Other arteries in the vicinity that may be damaged during surgery of the sphenoid sinus include the posterior ethmoidal artery, as well as the artery of the Vidian canal. Detailed knowledge of the nasal vascular anatomy is helpful in avoiding unwanted and severe operative bleeding.
Internal Carotid Artery Injury
The internal carotid artery is at risk during standard sinus surgery when there is excessive new bone formation and attempts are made to remove the new bone and when intersinus septations are taken down. In a large number of patients the intersinus septum in the sphenoid attaches to the carotid artery in the lateral wall of the sphenoid sinus. When the septum is taken down care should be taken that the septum is not grasped and rotated as part of the attempt to remove it. This technique can result in a sharp bony base of the septum fracturing, and during rotation of one of these, sharp bony spurs puncturing the carotid. The artery is also at risk if the wall covering the sphenoid protrusion of the carotid is dehiscent. This occurs in ~ 10% of patients and puts the vessel at risk if instruments or microdebriders are used in the sphenoid sinus. A much higher risk of carotid artery injury is present during endoscopic skull base surgery, especially if the tumor involves the carotid. This is commonly seen in pituitary tumors that extend into the cavernous sinus and in skull base tumors of the clivus such as chordomas and meningiomas that may involve the carotid.
If an injury should occur to the carotid, the following steps should be followed and are summarized in Fig. 9.9 . Help should be called for—both for the anesthetist and for the surgeon. A second surgeon will significantly improve the likelihood of a good outcome for the patient as endoscopic management of this situation usually requires two surgeons. The anesthetist should actively resuscitate the patient and keep the systolic pressure reasonable—this allows for continued contralateral blood flow from the opposite carotid through the cerebral circulation and helps to maintain cerebral perfusion. The surgeon needs to immediately harvest muscle from either the patient′s thigh or sternocleidomastoid muscle. This muscle should be ~ 1.5 by 1.5 by 1.5 cm. This muscle should be crushed—usually between two metal kidney dishes, usually available on the scrub nurse′s table. Two high-flow suctions are needed. The second surgeon places one suction down the side where the majority of bleeding appears to be coming from and tries to take as much of the blood flow away as possible. The primary surgeon now places his large volume suction and endoscope down the opposite nostril keeping the surgical field clear as the endoscope is advanced into the region of the bleeding vessel. The first surgeon then places his suction over the bleeding artery and hovers this suction directly above the site of injury. If his suction is great enough—and one should always be using the strongest possible suction—most of the blood should be attracted to and suctioned up. The primary surgeon now has the lesion in clear vision and can substitute his suction for the crushed muscle patch held by a Blakesley forceps. While the second surgeon keeps the blood flow away from the primary surgeon′s side, the primary surgeon slides the muscle patch directly on to the lesion keeping pressure on the patch and lesion during this maneuver. Continued oozing from the lesion is cleared by the second surgeon so that the muscle patch can be seen to be correctly placed on the lesion, the patch should at this stage be controlling the flow from the lesion.
The muscle patch should be held in place for at least 5 but preferably 10 minutes. The second surgeon can now bring a neurosurgical pattie onto the muscle patch and the Blakesley pressure can be slowly lessened. The second surgeon applies gentle pressure to the muscle patch by putting pressure on the pattie and the Blakesley forceps should be able to be withdrawn without the bleeding starting again. Now the pattie is gently removed and the bleeding should have stopped with the muscle patch. A few squares of Surgicel (Ethicon Inc,. Somerville, NJ, USA) are then placed over the patch and if the bleeder is in the sphenoid, a pedicled septal flap is rotated into the sphenoid to cover the muscle patch. This is glued into place and covered with Gelfoam and a pack (ribbon gauze or other) is placed over the flap to allow continued gentle pressure to be applied to the flap and muscle patch.
The patient is kept intubated and asleep and an immediate angiogram is performed to ensure that control has been achieved and to see if there is any ongoing ooze. If there is poor control or continued leakage then endovascular intervention is required and the vessel is either stented or coiled. The pack is then removed under general anesthesia 5 days later. If the initial angiogram was normal this should be repeated at 6 weeks and 3 months to ensure that no pseudoaneurysm has formed.
Intracranial Injury/Cerebrospinal Fluid Leak
The risk of CSF leak is a constant concern in ESS. Most series report a rate between 0.4 and 0.8%,2,13 although a recent nationwide audit of 40,638 ESS cases in the United States between 2003 and 2007 reported a rate of 0.17%.3 Transgression of the bone and dura of the skull base will result in a leak of CSF; this can usually be recognized at the time of surgery by a “washout” of clear fluid from the area of the injury, caused by a dilution of the blood covering the surrounding tissues. In instances where there is an abundance of inflamed tissues and bleeding around the site of injury, the CSF leak may look like a sudden onset of brisk venous bleeding, without any noticeable “washout”.2 A high index of suspicion for CSF leak must be maintained when any sudden increase in bleeding occurs in the vicinity of the skull base. If unrecognized or untreated, the leak of CSF can lead to postoperative pneumocephalus, tension pneumocephalus, meningitis, encephalitis, or epidural or subdural abscess.
Another risk of unrecognized skull base injury at the time of surgery is that of intracranial injury, including damage to cerebral vasculature or to the brain itself. The severity of the injury is dependent on several variables, such as the size and shape of the instrument involved, the type of instrument (powered debrider, electrocautery, cold steel), the depth of penetration, the time lapse between skull base penetration and recognition of the complication by the operating surgeon, and the anatomical structures injured. Depending on the structures affected, the sequelae of intracranial injury may include persistent headache, neurologic deficit, intracranial hemorrhage, and intracranial infection. Meningoencephalocele may occur in the late postoperative setting. Fortunately, these are exceedingly rare occurrences; the reported rates for major intracranial complications in ESS are 0.47 to 0.54%.2
A frequent location for iatrogenic CSF leak to occur is along the anterior vertical lamella of the fovea ethmoidalis constituting the lateral wall of the olfactory fossa. This is near the junction of the middle turbinate attachment and the cribriform plate ( Fig. 9.10 ). The bone in this area makes up the most medial aspect of the frontal recess dissection, and is the thinnest area of the skull base, measuring as little as 0.1 mm in thickness. It is also perforated by the anterior ethmoid artery. If damaged in this region, electrocautery of this vessel can lead to transmitted thermal injury of the skull base and dura causing an immediate or postoperative CSF leak. This risk is greater with monopolar cautery and can be minimized by using bipolar cautery. Damage to this area can also occur as a result of dissecting instruments being directed toward the olfactory fossa during dissection of the frontal recess. Hence, dissecting instruments such as curettes and probes should be maintained in an upright orientation and used to apply force in a posterior to anterior direction when fracturing bony septations in the frontal recess and along the skull base.
Endoscopic surgery involving the frontal sinus and its drainage pathway is among the most challenging aspects of ESS. As such, surgery in this critical area requires specialized training and expertise. The posterior table of the frontal bone marks the anterior limit of the skull base and anterior cranial fossa, as does the posterosuperior aspect of the frontal recess. The maneuvers in frontal sinus surgery most likely to cause a skull base injury include improper placement of dissecting instruments during the removal of obstructing frontoethmoidal cells in the frontal recess. In certain patients, the cell patterns contained within this area can be very complex and confusing, and it is easy for the unprepared surgeon to become disoriented intraoperatively.
Another relatively common area for skull base injury to occur intraoperatively is along the posterior ethmoid roof near the anterior face of the sphenoid sinus. This may occur if the surgeon is unsure of the position of the natural sphenoid ostium and attempts to force an instrument such as a Freer elevator, straight curette, or microdebrider into the ethmoid fovea, under the false impression that the skull base is located more superiorly ( Fig. 9.11 ). This can result in CSF leak or, if initially unrecognized, damage to intracranial structures like the brain, arterial blood vessels, or venous sinuses. Given the seriousness of the potential injuries to the skull base, optic nerve, or internal carotid artery, it is strongly advisable not to use powered instrumentation when entering the sphenoid sinus or enlarging the sphenoidotomy.
Alternatively, skull base injury may occur when bringing down the “frontal T” during the frontal drill out (endoscopic modified Lothrop, or Draf III) procedure. As one of the final steps in the drill out, the removal of bone anterior to the olfactory bulb is crucial to maximize the anterior–posterior dimension of the frontal neo-ostium and decrease the chance of postoperative restenosis; this step clearly defines the T-shaped anterior projection of the cribriform plate ( Fig. 9.12 ). The maximum posterior limit is achieved by identifying the first olfactory neuron that forms the anterior boundary of the olfactory fossa. Gradual drilling of this bone should be performed with the aid of image guidance to confirm the anterior limit of the skull base, and extreme care must be taken not to allow the drill to slip above this projection of bone, as the skull base is much thinner in this region.
A precise understanding of the anatomy in each individual patient is of utmost importance and preoperative imaging must be carefully reviewed before undertaking surgery. This should allow the surgeon to recognize any high-risk anatomical variants and give an appreciation of the likelihood of skull base injury during the dissection. High-risk variations include a low or asymmetric anterior skull base, a deep olfactory recess (Keros 2 or 3), a slanted lateral lamella of the olfactory groove that is tilted away from the vertical plane ( Fig. 9.10 ), as well as any expansile processes such as mucoceles or masses that have caused demineralization and bone loss along the skull base ( Fig. 9.13 ).
Careful review of the preoperative CT imaging is also essential to achieve a three-dimensional understanding of the frontal drainage pathway and to create a surgical plan for the frontal recess dissection. In this way, the dissecting instruments can be placed precisely along the elucidated pathway, atraumatically and with minimal resistance, to fracture away the obstructing frontoethmoidal cells in a stepwise fashion that was predetermined by the surgical plan.14,15 A key principle to safe dissection is that no dissecting instrument should ever be forced through the roof of a frontoethmoidal cell because of the risk of causing a CSF leak; rather, the instrument should be passed either medially or posteriorly to the cell wall according to the position of the frontal drainage pathway. In situations where there is excessive bleeding that is obscuring the surgeon′s view in this delicate area, suction instruments such as the Wormald malleable suction curette (Medtronic ENT, Jacksonville, FL, USA) are ideally suited for performing the dissection. The use of powered instruments such as the microdebrider should be minimized in the frontal recess, and if they are to be used, aggressive debriding should only be directed in an anterior direction toward the frontal beak, as this will avoid placing the skull base at risk. Finally, in cases of complex anatomy, exuberant disease, or previous surgery obscuring the drainage pathway intraoperatively, frontal sinus minitrephination (Medtronic ENT) can be helpful. By instilling fluorescein through frontal minitrephines, the frontal sinus drainage pathway is clearly highlighted by a stream of brightly colored fluorescein solution. Dissecting instruments can then be precisely placed into this corridor and the surrounding cells can be fractured away.
As mentioned previously, early intraoperative recognition and repair of a CSF leak is important in minimizing the likelihood of serious or long-term sequelae. However, it is also possible that an injury to the skull base goes unrecognized at the time of surgery and that the ensuing CSF leak is diagnosed during the postoperative period. The leak is often diagnosed on a clinical basis, with patients complaining of clear watery rhinorrhea that is salty to taste, and exacerbated on leaning forward or with a Valsalva maneuver. Biochemical confirmation can be achieved with a β2-transferrin assay; the presence of glucose in the fluid is also suggestive of CSF. Imaging studies that may prove useful include CT cisternography and radionucleotide scanning. Intraoperative localization can be facilitated by intrathecal instillation of fluorescein, which will color the CSF fluorescent green. Our protocol is to withdraw 10 mL of CSF from the lumbar puncture and add 0.1 mL of filtered 10% fluorescein, creating a 0.1% final concentration that is slowly infused intrathecally 30 to 60 minutes before the procedure.