Since the first attempts to remove pituitary tumors through the transsphenoidal approach almost a century ago, it became clear that one of the main challenges was to create an effective repair of the sellar floor to prevent postoperative cerebrospinal fluid (CSF) rhinorrhea, pneumocephalus, and meningitis. Traditional dural and bony closure techniques commonly used in transcranial neurosurgery such as primary dural suturing and bone flap fixation are generally not feasible in transsphenoidal surgery. Consequently, over recent decades, numerous sellar floor closure techniques have been devised using a variety of strategies and materials, both autologous and synthetic. In this chapter we provide a summary of the current state of closure techniques in transsphenoidal surgery preceded by a review of the frequency and degree of this intraoperative problem and the rate of associated postoperative complications.
Rate of Intraoperative and Postoperative CSF Leaks After Transsphenoidal Surgery
Intraoperative CSF Leaks
A CSF leak occurring at the time of transsphenoidal surgery has been reported to occur in 6% to 100% of procedures ( Table 15-1 ). In a surgical series of patients with pituitary adenomas, the rate of intraoperative leaks has ranged from 6% to 60%. , In contrast, in a series of extended transsphenoidal approaches for suprasellar tumors, such as craniopharyngiomas and tuberculum sellae meningiomas, the intraoperative CSF leak rate generally approaches 100% given the necessity of a large bony, dural and arachnoidal opening to access the tumor.
|Series||Rate of CSF Leak||Include Extended Approaches|
|Balagura et al, 1981||6%||No|
|Kabuto et al, 1998||23.2%||No|
|Elias & Laws, 2000||17%||No|
|Spaziante et al, 2000||10.2%||No|
|El-Banhawy et al, 2000||46.7%||No|
|Seiler & Mariani, 2000||18%||No|
|Kelly et al, 2001||32.25%||No|
|Cappabianca et al, 2002||14.1%||No|
|Zada et al, 2003||57%||Yes|
|Cappabianca et al, 2004||28.9%||No|
|Cook et al, 2004||100%||Yes|
|Esposito et al, 2004||54%||Yes|
Postoperative CSF Leaks
In most contemporary transsphenoidal surgical series for pituitary adenoma removal, the rate of postoperative CSF leak is well less than 10% and generally ranges up to 5.1%. In extended transsphenoidal series, the rate has been higher, ranging up to 21%. *
* References , , , .The associated complication of meningitis after transsphenoidal surgery is directly associated with the incidence of CSF leak and ranges from 0.5% to 14%. Tension pneumocephalus, a potentially devastating complication after transsphenoidal surgery, has been reported to occur at a very low rate of well less than 0.5%.
Principles of Transsphenoidal Sellar Defects and Their Repair
Cause and Size of CSF Leaks
Although at first glance, it may appear obvious how an intraoperative CSF leak may occur during transsphenoidal surgery, not all leaks occur in the same manner or to the same degree. Such CSF leaks generally occur in one of three ways. First, they may occur by simply removing the tumor that is directly against an already thinned and incompetent diaphragma sella. Second, they may occur as a result of traction associated with tumor removal, leading to a tear in the diaphragma sella. Third, they may result from a wide and often deliberate opening of the diaphragma to access lesions with supradiaphragmatic extensions or from a dural opening above the diaphragma sella directly into the suprasellar space. These three causes of an intraoperative CSF leak generally result in relatively small-, medium-, and large-sized leaks, respectively, although this correlation does not always hold true. However, it is clear there is a wide spectrum of CSF leak size and the approach to repairing these leaks can be tailored to the magnitude and location of the leak. To simplify this approach conceptually and practically we use a CSF leak grading system that includes grade 0—no CSF leak, and grade 1 to grade 3 CSF leaks ( Table 15-2 ). A grade 1 CSF leak is characterized by a small “weeping” leak without an obvious rent or only a small rent in the diaphragma sella but that is clearly evident spontaneously or when a Valsalva maneuver is applied. A grade 2 CSF leak is characterized by a well visualized medium-sized rent in the diaphragma sella in which CSF can be seen easily flowing into the intrasellar space. A grade 3 CSF leak is characterized by a large and often deliberate diaphragmatic opening or an opening of the tuberculum sella dura above the diaphragma, typically as part of an extended transsphenoidal approach for a suprasellar tumor.
|Grade 0||Absence of CSF leak, confirmed by Valsalva maneuver|
|Grade 1||Small “weeping” leak, confirmed by Valsalva maneuver, without obvious or only small arachnoidal defect detected|
|Grade 2||Moderate CSF leak, with obvious arachnoidal defect|
|Grade 3||Large CSF leak, typically created as part of extended transsphenoidal approach through the supradiaphragmatic or clival dura for tumor access|
Key Principles of Sellar Repair
There are two principles that must be followed to achieve successful repair of a sellar CSF leak, regardless of the size of the leak. First, the material used to block the egress of CSF, whether it be fat, muscle, fascia, or synthetic collagen, must provide an effective water-tight barrier to the flow of CSF. Second, a solid or semisolid buttress must hold the sealing tissue in position and maintain the water-tight seal to resist the pulsations of the brain and CSF. This buttress, whether it is septal cartilage or bone, or a synthetic plate or titanium mesh, is best positioned in the intrasellar extradural space.
There are two other important details to consider that can further add to the effectiveness of a given repair and reduce the likelihood of failure. These considerations are whether to (1) use tissue glue, such as fibrin glue, Tisseel (Baxter, Deerfield, Ill.) or BioGlue (CryoLife International, Inc.; Kennesaw, Ga.) to help further ensure a water-tight repair ; and (2) insert a lumbar drain for CSF diversion to further decompress the repair. The use of these additional tactics should be dictated by the size of a particular CSF leak and the perceived effectiveness of the first two steps of the repair.
An additional requirement for an effective repair is that of achieving adequate hemostasis after tumor removal. This goal is best obtained by ensuring as complete a tumor removal as possible. For sellar dural bleeding, bipolar coagulation of obvious bleeding sites is often effective. However, the bipolar should be used with caution and at a low power setting given that extensive cauterization can cause dural retraction and enlargement of an existing dural/diaphragmatic defect or can precipitate a CSF leak de novo where the diaphragma attaches to the sellar dura. To avoid enlarging an existing leak by cauterization, particularly for diffuse intrasellar oozing that often occurs after removal of a large macroadenoma, use of hemostatic agents such as Gelfoam, Surgifoam (Johnson & Johnson), and Surgicel (Johnson & Johnson) are often effective when placed in the resection cavity and held in position with a cottonoid for several minutes.
Summary of Currently Used Closure Materials
Over the last 3 decades of the microsurgical era in transsphenoidal surgery, a variety of repair techniques using a wide spectrum of materials have been described. Probably the most widely used sealing materials are autologous fat harvested from the abdomen or thigh and fascia lata harvested from the thigh. Occasionally muscle has also been used. Fascia and muscle have also been used in combination with fat placed in the sphenoid sinus to provide a buttress effect without addition of a more rigid buttress. Many pituitary surgeons, however, use a rigid buttress to hold the sealing material (fat or fascia) firmly in place. These materials include septal cartilage and bone or sphenoid bone and middle turbinate mucoperichondrium. The use of such autologous tissue is advantageous from the standpoint of being biocompatible and “free of charge” to the patient. The disadvantage of harvesting fat, fascia, or muscle is that a separate surgical incision is required. Abdominal or thigh incisions generally heal very well with a low complication rate, but they can be associated with wound infections, hematoma, incisional pain, and cosmetic defects.
Allografts and Synthetics
More recently, a variety of allografts and synthetic materials have been introduced into the repertoire of sellar repair techniques for both the sealing material and the buttress material. The rationale for using such materials is that they obviate the need for additional incisions and can provide an effective repair barrier without autologous grafts. Furthermore, in cases of endonasal endoscopic or endonasal microscopic approaches, which are increasingly common, the bony and cartilaginous nasal septum is not disturbed by the approach, other than being mobilized, so harvesting septal tissue is avoided.
The material currently used most frequently as a substitute for dura and in lieu of (or with) abdominal fat and fascia is collagen sponge, typically prepared from bovine flexor tendon. This collagen matrix (DuraGen [Integra, Plainsboro, N.J.]; Helistat [Integra]; Instat [Johnson & Johnson, Somerville, N.J.]) is effective because it acts as a scaffolding for rapid fibroblast in-growth and ultimately becomes a protective water-tight barrier over the sellar contents. As described below, in most instances of small (grade 1) weeping CSF leaks, a two-layered collagen sponge repair with an effective buttress can obviate the need for autologous tissue grafts, tissue glue, and lumbar CSF diversion. However, as described below, for larger CSF leaks (grade 2 and 3 leaks), it is recommended that an autologous tissue graft, preferably fat, be used to ensure a successful repair. An additional advantage of collagen sponge is that it stimulates factors VIII and XII of the coagulation cascade, thereby promoting hemostasis, and it promotes the adhesion and aggregation of platelets.
Synthetic Buttress Materials
Numerous synthetic materials have been used as a sellar buttress in transsphenoidal surgery, including titanium plates and mesh, nonresorbable plates made of ceramic, silicone, polyester-silicone and polyethylene, and bioabsorbable patches and plates. *
* References , , , , .The three essential properties of sellar buttress materials are that they be (1) rigid enough to hold the sealing material (fat, fascia, muscle, or collagen) in place and resist intracranial pulsations; (2) malleable enough so that they can be tailored to the specifics of a given bony and dural defect; and (3) relatively inert with a low propensity for infection. It is also helpful if the buttress is stable in size once deployed inside the sella and easily recognizable on postoperative imaging studies including CT and MRI. Although stainless steel has been used in the past, it would appear that it is infrequently used at this time given that it is not MRI-compatible.
As shown in Table 15-3 , in recent years, the most commonly used buttress materials appear to include (1) nonresorbable synthetic plates and sheets made of silicone, polyester-silicone, and polyethylene ; (2) bioabsorbable synthetics including the amorphous polymers that become malleable at elevated temperatures and the Vicryl patch ; and (3) titanium mesh. In most of these reports, when CSF leaks are present, additional materials are used in repair, most typically fat grafts, and in some instances tissue glue and lumbar CSF diversion.
|Resorbable Vicryl patch||Minimal||Yes||Yes||Yes|
|Amorphous polymers ( MacroSorb )||Yes||Yes *||Yes||Yes|
|Silicone plate||Yes||Yes||Rare FB reaction †||Yes|
|Polyethylene plate (Medpor)||Yes||Yes *||Yes||Yes|
Of the nonresorbable synthetics, all appear to have adequate rigidity, are easily cut with scissors, and are relatively inert with a low risk of infection or foreign body reaction. In particular, silicone plates appear to provide an effective buttress and are easily detected on MRIs with a very low rate of delayed foreign body response. Similarly, a polyester-silicone dural substitute has been effectively used in purely endoscopic approaches. The material is introduced in bent fashion through the nostril, to be released and spontaneously opened in the intradural or extradural space where a reactive fibrosis occurs thereafter. Its elastic properties allow easy and safe passage through the nasal cavity without mucosal trauma. It too is easily seen on CT and MRI and easily identified and incised at reoperation. Another similar product recently introduced is Medpor (Porex Corp., Fairburn, Ga.), a porous, high-density, 0.4 mm polyethylene sheet that can be easily shaped in hot water before insertion as a buttress within the sella or the sphenoid sinus.
Among the bioabsorbable substances, amorphous polymers are being increasingly used for sellar buttress repair. In particular, MacroSorb (MacroPore, San Diego) is a polylactide polymer that is malleable at temperatures of 70° C and regains rigidity rapidly at body temperature. It has been used effectively for sellar repair where it can be easily heated, contoured, and placed within the bony defect. The plates are resorbed within 18 months of implantation. Another bioabsorbable product is the Vicryl patch, which when used in conjunction with fibrin glue had a low failure rate of only 0.05% in one report, despite the Vicryl being less rigid than most of the other synthetic buttress materials described here.
Of all metals available, titanium is the one used currently with the highest frequency in neurosurgery given that it is MRI-compatible. For transsphenoidal surgery, a thin titanium mesh (Micro Mesh 0.2 mm, Stryker Leibinger, Kalamazoo, Mich.) is ideally suited as a sellar buttress, being highly malleable, easily cut with scissors but relatively rigid and totally inert. The details of using titanium mesh for sellar floor repair are described later.
Closure Technique: Graded Repair of Intraoperative CSF Leaks
In this section, a detailed description of the four types of sellar floor repairs is provided using the principles and concepts described previously ( Table 15-4 ). This approach is based on our experience with 437 consecutive direct endonasal microscopic cases for tumor removal in the sellar and parasellar area. As noted above, the specific repair methods outlined are dictated largely by the size of the intraoperative bony and dural defects. Nasal packing is not used regardless of the size of the leak.
|Grade of Leak||Repair Method|