SPECIALIZATION IN OTORHINOLARYNGOLOGY
Endoscopic repair of cerebrospinal fluid leaks: systematic
literature analysis.
Author: Supervisor:
Dr. Alessandra Iavarone Dr. Giuditta Mannelli
Table of contest:
1. Introduction
2. Etiology of CSF Leak
3. Diagnosis of CSF Leak
3.1 Clinical presentation and Imaging
3.2 Intrathecal fluorescin
4. Surgical Repair of Cerebrospinal Fluid Leaks:Local, Regional and
distant flaps
5. Endoscopic approaches
6.
Reconstructive techniques
6.1Nasoseptal Flap (NSF)
6.2 Inferior Turbinate Flap (ITF)
6.3 Posterior Pedicled Middle Turbinate Flap
6.4 Three-layer reconstruction with the Iliotibial Tract (TRITT)
6.5 Inlay-Onlay techniques
6.6 Gasket seal closure
6.7 Bath plug technique
7. Perioperative managment
7.2 Lumbar drain
7.3 Antiobiotcs
8. Success rate
9. Complications
10. Objective of the review
11. Materials and methods
11.1 Study characteristics and quality assessment
11.2 Statistical analysis
12. Results
12.1) Demographics
12.2) Factors relating to CSF leak
12.3) Factors relating to adjuvant management
12.4) Factors relating to repair
12.5) Factors relating to post-operative period
13. Discussion
14. Conclusion
1. Introduction
The anterior skull base is a key barrier between the intracranial intradural compartment on the one hand and the sinu-nasal tract on the other. It is inevitable for the prevention of ascending infection, loss of cerebrospinal fluid (CSF) or brain prolapse.
Irrespectively of their idiopathic, traumatic, postoperative or tumorous nature, defects of anterior skull base have to be closed meticulously.
Historically, the closure of these defects was approached intracranially using classic open skull base surgical procedures. Although the open intracranial approach provides a large surgical field and allows for direct visualization, it is associated with high morbidity, including intracerebral hemorrhage, cerebral edema, frontal lobe deficits, lengthened hospital stay, anosmia, and high recurrence rates.
Over the past fifty years, CSF closure techniques have undergone significant evolution. Minimally invasive approaches, characterized by improved success rates and decreased morbidity, have gained increasing favor over the traditional open craniotomy repairs1.
Dandy described the first successful intracranial repair of a CSF leak through a bifrontal craniotomy in 1926; this remained the procedure of choice until the late 1940s. The disadvantages of a frontal craniotomy were the substantial morbidity of the craniotomy and the permanent anosmia from mobilization of the olfactory bulbs. Despite the magnitude of the procedure, closure of the leak is not guaranteed after the craniotomy approach. The recurrence rate of CSF leaks has been reported to be as high as 27% after the first attempt, and 10% of patients have persistent leaks despite multiple attempts. Park et al reported successful closure of CSF rhinorrhea after craniotomy in only 12 of 20 (60%) patients.
A second extracranial approach for the repair of CSF rhinorrhea was reported in 1948 by Dohlman, who described a naso-orbital incision. The external ethmoidectomy approach was recommended by Chandler and is the method most commonly used when there are no intracranial complications requiring management.
Transnasal closure of the sphenoid CSF leaks was reported by Hirsch in 1952. Vrabec and Hallberg described a cribriform plate leak repair using an intranasal approach, but they needed a simultaneous submucous resection of the nasal septum for adequate visualization. The defect was repaired with an advancement flap from the turbinate. Lehrer and Deutsch reported an additional two cases and suggested the use of the operating microscope. The techniques described in all of these papers used a combination of head-light illumination and the operating microscope2 .
Advancements in endoscopic technology and improved surgical techniques have led to the development of the transnasal endoscopic approaches for CSF leaks. Endoscopic closure was first mentioned in a paper on endoscopic ethmoidectomy by Wigand in 19813 and a small series was
subsequently published by Mattox and Kennedy in 19904.Subsequently, endoscopic repair has
become an accepted standard of care for the operative management of most CSF leaks. Multiple publications site the low incidence of complications and low rates of recurrence associated with a minimally invasive endoscopic approach5.
2. Etiology
Cerebrospinal fluid (CSF) rhinorrhea denotes a skull base fistula connecting the subarachnoid space to the nasal cavity. Defects commonly occur in the ethmoid roof, the cribriform plate, and the sphenoid, and less commonly in the frontal sinus posterior table.
In 2004, Schlosser and Bolger6 published a review based on the etiologic factors associated with
CSF rhinorrhea, discussing accidental trauma, surgical trauma, tumors, congenital and spontaneous sources.
Lopatin et al.7 classified them into primary (spontaneous) and secondary CSF rhinorrhea, whereas
Gendeh et al.8 suggested only three etiologic categories: congenital, acquired, and spontaneous.
As Hegazy et al9 described, accidental trauma being the most frequent cause (44%), followed by
surgical trauma (29%) and tumors (22%) . Spontaneous and congenital causes occur as well.
Most commonly, the leak is found at a cribiform plate location (35%), followed by sphenoid sinus (26%), anterior ethmoid (18%), frontal sinus (10%), posterior ethmoid (9%) and inferior clivus (2%) 10.
CSF leak in the setting of accidental trauma has been associated with both a high rate of spontaneous closure (70%) and an elevated risk of ascending meningitis (30 – 40%)11.
Posttraumatic leaks were more common in males aged 20-39 years. Locatelli , in a 2006 series, reported the cribiform plate (23.1%) and ethmoid skull base (20.5%) as the most common sites in fracture, with 35.9% having multiple sites of injury.
Iatrogenic CSF leaks occurs as the result of endoscopic sinus surgery and neurosurgical procedures. Surgical lesions can range from small dural rents to larger skull-base resections. Iatrogenic injury
most commonly occurs at ethmoid skull base (35.1%), the cribiform (27%) and the sphenoid sinus (18.9%) 11.
Several circumstances predispose the patient to have a CSF fistula during endoscopic sinus surgery. Failure to appreciate the anatomical relationships, because of the surgeon's lack of experience or ill-defined anatomical landmarks caused by bleeding, the presence of space-occupying lesions, previous surgery, or anatomical variations, is the most important factor leading to CSF leak. Most authors agree, however, that previous endoscopic surgery (revision) constitutes the most common clinical scenario associated with iatrogenic CSF fistulae 9.
Congenital skull base defects are generally rare, and 63% of them occur in the foramen caecum12.
Spontaneous CSF rhinorrhea remain a diagnostic and surgical challenge. The frequency of spontaneous leaks has been reported10as between 15% and 23%. The cribriform plate is a common site, although Gendeh et al.8 found 40% to 56% of spontaneous leaks in the sphenoid. Spontaneous
leaks were more common in female aged 40-59 years.
Most recent research13,14 has focused on the role of intracranial pressure in both the origination of the leak and in the increased failure rate associated with spontaneous leaks. High intracranial pressure (ICP) is associated with 63% to 88% of spontaneous leaks.
Spontaneous CSF leaks are associated with a 50% to 100% incidence of encephalocele and a 25% to 87% incidence of recurrence6.
The etiology (Fig.1) of spontaneous CSF (sCFS) leaks is not completely understood, but there is a clear association of sCSF leaks with obesity (80% of patients), elevated ICP (40% of patients) and obstructive sleep apnea OSA (43% of patients). Data from the National Health and Nutrition Examination Survey (NHANES) have demonstrated high rates of obesity in the United States beginning in the early 1980s15.
It is currently estimated that 35.2% of males and 40.5% of females are obese. Recent NHANES data demonstrate a stabilization of the obesity rate in the United States for men since 2005, but a continued rise in the obesity rate for women16.
There is also a strong association with obstructive sleep apnea (OSA) in patients with sCSF leaks. As many as 43% of patients with sCSF leak presented with the diagnosis of OSA and the incidence
of OSA may be higher if all patients with sCSF leaks were prospectively tested for OSA17.
This association is important because it is know that ICP spike during apneic events suggesting that
episodic rises in ICP may also contribute to skull base erosion over time18. Thus, it is recommended
that all patients with sCSF leaks undergo a polysomnogram to assess for OSA. After surgical repair
of lateral skull base sCSF leaks, it appears safe to resume CPAP treatment of OSA19.
It is unknown if CPAP treatment of OSA can delay or prevent the occurrence of sCSF leaks.
Weight loss should be encouraged but currently there is no data showing that weight loss or bariatric surgery can alter ICP or the incidence of sCSF leaks.
Fig.1. Surgical Repair of Spontaneous Cerebrospinal Fluid (CSF) Leaks: A Systematic Review Brian C. Lobo, MD;
Spontaneous CSF leaks have the highest recurrence rate following surgical repair (25–87%),
compared with <10% for most other etiologies6,21. In a retrospective analysis of 72 patients over a
10-years period, Mirza et al., observed that 13 of 29 patients with spontaneous CSF leaks (46%) had evidence of raised intracranial pressure; 6 of the 13 patients with raised intracranial pressure (46%)
had a recurrence of leak22.
Chaaban et al., on the basis of their 5-years prospective study on 46 patients with 56 spontaneous CSF leaks, concluded that successful treatment of elevated intracranial pressure in combination with endoscopic repair can provide high success rates (93% primary and 100% secondary) approaching that of other etiologies. 23
A high incidence of recurrence also characterizes CSF rhinorrhea due to neoplasms or to very large defects.
3. Diagnosis
3.1 Clinical presentation and Imaging
Among frequent symptoms one may find a watery rhinorrhea, mainly unilateral, and sometimes headaches when the fistula is associated with a meningocele or ascending meningitis24.
Inherent in successfully treating CSF rhinorrhea is accurate diagnosis of the presence of a leak and precise localization of the leak. The presence of CSF in clear nasal drainage should be established through analysis for CSF markers. The most commonly used is the assay for beta-2 transferrin. The
use of this marker has been well established as both a sensitive (97%) and a specific (93%) assay for the presence of CSF13. In 2004, Schnabel25 compared the specificity and sensitivity of beta-2
transferrin to beta trace protein, finding both to be effective, with beta trace protein having a slightly improved sensitivity (100%) and specificity (100%).
In a literature review of 39 papers on the utility of testing beta-trace protein or beta-2 transferrin Bachmann-Harildstad could show that Beta-trace results being faster (20 minutes versus 120 minutes of beta- 2 transferrin) and less expensive26.
Localization of the leak site has evolved and radiologic techniques have advanced. Numerous authors recommend 1-mm thickness axial and coronal computed tomography (CT) scan with bone algorithm14.Additional localization techniques include CT cisternogram and magnetic resonance
(MR) imaging.
CT cisternography utilized intrathecal water-soluble iodine contrast material in place of the formerly used metrizamide. Its sensitivity ranges from 48% to 96% depending on the flow rate and site. However, it can miss intermittent leaks. It is contraindicated with high ICP and holds low acceptability. Its interpretation is operator dependent27. MR imaging using T2-weighted images and
MR cisternography (a noninvasive, fast-spin echo protocol) can also aid in localization.
Magnetic resonance imaging (MRI) cisternogram (T2 or fast spin echo with fat saturation and image reversal) can also miss intermittent leaks. Its accuracy is 89%, sensitivity 85% to 92%, and specificity 100%. However, Hegarty and Millar 28 reported up to 42% false positives with this
method.
Image guidance systems utilizing multiple planar CT information have become frequently used tools in endoscopic sinus surgery. Their role in the management of CSF leaks has not, to date, been widely studied. In 2005, Tabaee29 reported a retrospective cohort of patients with CSF rhinorrhea,
finding that the image guidance system improved the confidence of the surgeon but had no bearing on surgical success rates.
Fig.2 Algoritm of the management of CSF leaks. (CISS = constructive interference in steady state.)
Management of Anterior Skull Base Defect Depending on Its Size and Location.Manuel Bernal-Sprekelsen,Elena Rioja, Joaquim Enseñat,Karla Enriquez, Liza Viscovich,Freddy Enrique Agredo-Lemos,and Isam AlobidBioMed Research International Volume 2014 30
3.2
Fluorescin
The use of sodic fluorescein to detect a suspected CSF leak started back in 1960 to impart its fluorescence properties to CSF and facilitate its identification 31.
The use of this substance for diagnostic and surgical means purposes gained consent by proving to be effective at low doses, specially in cases of small defects or multiple fistulae, when mistakes are most common.
Based on these considerations, intrathecal fluorescein is, for most Authors, the clinical approach of choice to CSF leaks. For other Authors, use of fluorescein might be a valuable aid but should be avoided due to possible complications.
Off label intrathecal use is neither indicated nor prohibited by the United States’ Food and Drug Administration. The Authors are not aware of explicit authorization for intrathecal administration of fluorescein, in any country in the world, although the procedure is widely practiced worldwide, outside of the United States.
In an enquiry performed among rhinologists the habitual amount of fluorescein used was 0.5 and 1.0 mL at a concentration of 10%, although Senior et al.32 could show effectiveness at low
concentration and dosage of 0.1 mL at 10%.
Complications after intrathecal fluorescein injection are usually related to increased dosages or concentrations or to a high administration speed33. Severe complications (seizure, opisthotonus, and
peripheric palsy) have been linked to chemical irritation or overdosage, as could be shown by Syms III et al. in an experimental study.
From a review of the literature and the preliminary results of the Italian multicentre study, it can be seen that lumbar intrathecal fluorescein administration is a safe procedure provided that the following criteria are respected: maximal dose of 50 mg of fluorescein, further dilution of the colorant in CSF, and slow administration. With the exception of sporadiccases, the literature clearly shows the usefulness of the test, both in the diagnostic and intra-operative phase, as well as for preventive purposes, when surgical procedures on the skull base are performed with risk of dural damage. Many Authors have demonstrated that the off-label use of fluorescein, administered intrathecally, can be performed safely, in hundreds of patients, without any morbidity and mortality provided the correct solution is used, properly diluted, and carefully administered34.
4. Surgical Repair of Cerebrospinal Fluid Leaks:Local, Regional and
distant flaps.
Chakravarthi et al 37 outline a pragmatic algorithmic approach focused on vascularized flap
reconstruction that provides a general overview of possible reconstructive options. There are three main categories: local, regional and distant free flaps.
Local:
Middle turbinate flap (MTF) Inferior turbinate flap (ITF) Nasoseptal flap (NSF)
!
Fig .3 Anatomic coverage of reconstructive flaps. (A) Sagittal plane depicting the approximate coverage of the NSF, MTF, and ITF. The NSF can adequately cover the cribriform, planum, and upper two-thirds of the clivus. The MTF and ITF can cover small defects of the cribriform and planum and upper one-third of the clivus, respectively. 37
Regional:
Midline: pericranial flap (PCF)
Paramedian: temporoparietal fascial flap (TPF)
!
Fig.4 (B) The coverage of the PCF is the coronal midline plane. For paramedian defects, the TPFF is preferred. 37
Distant:
Anterior lateral thigh (ALT) flap Radial forearm flap (RFF)
CSF leaks managed via a craniotomy show 70-80% success closure rate. Advantages of the transcranial approach include direct visualization of the dural defect, the ability to address associated brain injury and the potential to use a large flap. However many studies have reported a 40% recurrence rate with this approach and significant patient morbidity including anosmia, frontal lobe retraction, seizures, memory deficits, and intracranial hemorrhage38.
In attempt to avoid these complications and improve closure rates, the endonasal endoscopic approach has evolved to address CSF leaks of the anterior skull base. The endoscope not only
provides excellent visualization, but outcomes studies have demonstrated decreased morbidity and improved closure rates exceeding 90% of success rate.
Other advantages of endoscopic approaches are preservation of smell, shorter hospital stay, absence of external scarring or brain retraction, and low risk of bleeding, seizures, and infections (<1%).
The approach also allows for close postoperative surveillance of the wound site and the ability to identify recurrences.
In a recent review Clavenna et al39 suggested the use of current pedicle flaps for defects more than 1
cm and for high-flow CSF leaks and asserted that the use of vascularized pedicled flaps in endoscopic skull base surgery has reduced CSF leak rates to well below 10%, comparable to that of open surgical techniques.
Analogous to the pericranial flap that is commonly used in open approaches, the nasoseptal flap (NSF), first described in 200640 quickly became the workhorse for repair of skull base defects
following endoscopic tumor surgery39.
Intranasal and Regional Vascular Flaps Available for Skull Base Reconstruction (Tab. 2)
Tab.2 How to Choose? Endoscopic Skull Base Reconstructive Options and Limitations Mihir R. Patel, M.D.,Michael E.
Stadler, M.D.,Carl H. Snyderman, M.D.,Ricardo L. Carrau, M.D.,Amin B. Kassam, M.D.,Anand V. Germanwala,
5. Endoscopic approaches
For all surgical approaches, standard preoperative preparation with topical nasal decongestants and injection of lidocaine and epinephrine will improve visualization and reduce hemostasis. The choice of graft is dictated by the size and the site of the defect.
Endoscopic approaches.
-Cribiform plate
To enable appropriate exposure of cribiform and ethmoid roof leak sites, anterior and posterior ethmoidectomy is performed. The remainder of natural ostial pathways of the maxillary, sphenoid and frontal sinuses should be addressed only to improve exposure or to ensure postoperative patency. Middle turbinate removal should be performed under the same guidelines.
-Sphenoid sinus and posterior ethmoid, including lateral pterygoid approach
CSF leaks localized to the sphenoid sinus must be further isolated to their precise location. Leaks located in the central portion of the sinus can usually be approached through standard transethmoid endoscopic approach, with possible conversion to transeptal assistance, depending on visualization.
Lateral sphenoid sinus leaks are often difficult to visualize and access safely by transethmoid approach. Bolger35 has described the endoscopic transpterygoid approach with anterior and
posterior ethmoidectomy and maxillary antrostomy. The dissection then proceeds with removal of the posterior face of the maxillary sinus to access the pterygopalatine fossa. The contents of the fossa are addressed by displacement and preservation if possible, including the internal maxillary artery, maxillary branch of the trigeminal nerve, the vidian nerve and the sphenopalatine ganglion. The sphenoid sinus is entered through this region at its anterior wall.
-Frontal sinus
Frontal sinus techniques, like sphenoid sinus techniques, depend solely on the location of the leak. Lateral and superior leaks may necessitate extracranial approach with osteoplastic flap and obliteration. Inferiorly based leaks, at the interface of the frontal recess and the cribiform, may benefit from combined endoscopic frontal approach. Leak sites encompassing the frontal recess and those near the native frontal outflow tract may be amenable to endoscopic techniques. Woodworth36
reported good success rates with comprehensive dissection of the frontal recess, with removal of agger nasi, frontal cells and suprabullar cells to provide adequate exposure. Maximal preservation of the mucosal surfaces has been stressed in numerous reports to avoid frontal recess stenosis. Individual patient anatomic factors play a large role in the potential success of this approach and need to be considered in the management paradigm.
6.
Reconstructive technique
6.1 Nasoseptal Flap (NSF)
The NSF is harvested as a mucoperichondrial flap based on the posterior nasoseptal artery. In general, an incision is made along the floor of the nasal cavity from the choanae to 1.5 cm posterior to the caudal septum with needle-point electrocautery.
The anterior incision is then made in a superior direction 1.5 cm posterior to the caudal septum, but the length can be tailored to the defect size. The superior incision is made with microscissors and carried back to the sphenoid face 1.0 to 1.5 cm below the skull base. The natural ostium of the sphenoid is identified. The mucosa is dissected off the anterior sphenoid wall creating a pedicle of mucosa from the sphenoid ostia to choanae and taken back to the level of the sphenopalatine foramen. This pedicled NSF is then placed in the nasopharynx or the maxillary antrum for the duration of the case. The defect is closed with a layer of collagen matrix (Duragen) inserted deep to the dura. A fat graft is then applied and held in place with a fibrin sealant (Confluent Surgical, Inc., Waltham, MA), and then the NSF is laid over it and secured with DuraSeal (Confluent Surgical Inc.). Septal splints are placed over denuded septal cartilage and bone and left in place for 3 to 4 weeks. The wound is then covered with a thick layer consisting of pieces of gel foam and then a Foley catheter balloon is used to support the flap and is inflated with 10 cc of saline. Two merocel sponges are often used underneath the balloon.
Fig.5 Pedicled NSF. (A) Lateralizing and outfracturing the middle turbinate provide an unhindered view of the nasal cavity. Then, superior and inferior incisions are made. The superior incision is made at the level of the sphenoid ostium, and the inferior incision is made at the level of the posterior choana. (B) As part of an extended flap, a long inferior incision is made at the level of the posterior choana. (C) After NSF is lifted, it is pushed into the nasopharynx, with a resulting denuded nasal septum and floor. (D) Contralateral septal mucosa following posterior septectomy. (E) Retracted NSF in relation to the crista ethmoidalis, sphenoid ostium, and anterior wall of sphenoid sinus. 37
Strengths of the NSF include its consistent vascularity, fairly long and robust pedicle, ease of harvesting, and the ability to customize its size and shape to cover small, large, and irregular ASB defects. The NSF can potentially cover defects encompassing up to 50% of the ASB, with the added
ability to span from orbit to orbit42,43. Furthermore, the NSF has a reported 90–95% success rate in
repairing ASB defects 44, a significant improvement compared with free grafts.
Challenges in extending the NSF to cover far anterior defects have limited its use in complex frontal sinus work, including lesions accessed via a modified Lothrop approach, and some anterior
cribriform closures.
Other shortcomings of the NSF include the resulting large anterior mucosal defect, which can result in crusting and possible nasal valve stenosis 45. This can be partially mediated by placement of a
free graft (middle turbinate mucosa often used) over exposed bone or cartilage to help promote healing at the donor site 46.
6.2 Inferior Turbinate Flap (ITF)
Indications : Due to the limited arc of rotation for the ITF compared with the NSF, its application is typically restricted to posterior defects of the sella or parasellar and midclival areas. The donor site can also result in prolonged crusting over the exposed inferior turbinate bone.
Contraindications: Contraindications are similar to other considerations, such as tumor invasion, previous surgery, and limited reach.
Surgical Technique: The vascularization of the ITF is based on the inferior turbinate artery, which is a branch of the posterior lateral nasal artery (a branch of the sphenopalatine artery). The posterior lateral nasal artery first runs along the inferolateral segment of the anterior process of the palatine bone and provides a branch to the middle turbinate medially. As the artery travels anteriorly, it becomes larger, connecting with the angular artery and supplying the anterior blood supply to the inferior turbinate.It is important to consider that certain anatomic variations exist in the blood supply of the sphenopalatine artery, one of which is that the posterior lateral nasal artery may extend anteriorly to the posterior wall of the maxillary antrum. This variant in particular is important to recognize in avoiding injury to the vascular pedicle when performing the maxillary antrostomy and mucoperiosteal elevation.
The first step of ITF harvest is medialization of the inferior turbinate to allow for visualization of the medial surface of the inferior turbinate. To visualize the natural ostium of the maxillary sinus, the uncinate process and the bulla ethmoidalis are then removed. Next, the maxillary sinus ostium should be widened toward the posterior maxillary antrum. With either a Cottle or Freer elevator, the submucoperiosteal mucosa overlying the anterior aspect of the ascending process of the palatine bone is lifted in a posterior direction toward the crista ethmoidalis, sphenopalatine artery, and foramen. Next, in a posterior to anterior direction, two parallel incisions are made to create the superior and inferior margins of the ITF. The superior incision starts at the posterior margin of the middle meatus at the site of the middle meatal antrostomy and extends superiorly to include the posterolateral margin of the nasal wall, along the nasal aperture, with the posterior limit being the Eustachian tube.
The inferior incision is a midline cut along the nasal septum, posteriorly extending to the Eustachian tube. During this step, it is important to sharply dissect the mucoperiosteum around the nasolacrimal duct to prevent its disruption. Next, a vertical incision at the head of the inferior turbinate, at its attachment to the piriform aperture, joins the superior and inferior incisions. Using either a Freer or Cottle elevator, the mucoperiosteum can now be raised from the floor of the inferior turbinate in an anterior to posterior direction and from medial to lateral. At this point, the tedious task of removing the internal cartilage without disrupting the overlying mucosa is now pursued. The pedicled and elevated flap often retains the shape of the inferior turbinate. Much of this is due to the presence of the midline ridge of the inferior turbinate overlying the mucosa. Therefore, the difficult step is to flatten the flap so that it fits securely on the defect.As a result,this bone is removed once the flap is elevated. The flap is then rotated into the skull base defect and set into place. Sometimes, a small Y-shaped cut has to be made in the distal end of the flap to allow for the flap to be placed in a flat position. It is notable that it can be challenging raising the flap after the turbinate bone has been fractured.40
6.3 Posterior pedicled middle turbinate flap
The posterior pedicled middle turbinate flap derives its blood supply from a posterior-based branch of the SPA. Limitations of this flap include the technical difficulty of elevating the flap from under- lying turbinate bone, particularly when anatomical variations in size and shape are present. The soft tissue component of this flap is also significantly thinner than the NSF. This option may be a consideration for small posterior ethmoid roof defects; however, a free mucosal graft may be more desirable in this setting because it can be precisely placed without limitation because of a pedicle.40
6.4 Three-layer reconstruction with the iliotibial tract (TRITT)
Three-layer reconstruction with the iliotibial tract (TRITT) is a safe and reliable alternative when vascularized flaps are not available.47
Surgical technique: The iliotibial tract graft is harvested on the lateral aspect of the thigh and divided in three portions, which are positioned in a multi-layered fashion to close the skull base defect. The first layer is disposed intracranial intradural, the second intracranial extradural, and the third extracranial. Fat grafts from thigh subcutaneous tissue are placed between the second and third layer to fill the dead space between them. Use of fibrin glue and intradural irrigation may help the surgeon to stabilize the layers during reconstruction. Step 1: Harvesting Of Itt. A 10-cm incision is performed in the middle third of the lateral aspect of the thigh, approximately 4 fingers below the line passing through the anterior-superior iliac spine and the lateral margin of the patella . Subcutaneous fat, thinner in elderly and males, is dissected to expose the whitish ITT. A rectangular area of ITT (about 10 x 6
cm) is taken together with pieces of subcutaneous fat. During the harvesting of ITT some tips should be kept in mind:
- The cranio-caudal orientation of its fibers makes the ITT very resistant to longitudinal forces, while transversal traction may tear it. This is particularly true in the elderly, where the ITT is thinner and more fragile.
- During dissection, any perforating vessel should be carefully cauterized.
- The graft is more easily cleaned from the fibro-adipose tissue that is adherent to the fascia before it is harvested.
- The cranial cut of the graft should not injure the tensor fasciae latae muscle in order to reduce the risk of postoperative bleeding.
- Any laceration of the ITT (iatrogenic or determined by the presence of perforating blood vessels) should be sutured with re-absorbable 4.0 stitches.
pliability.
The surgical wound should be meticulously sutured with a double line of subcutaneous stitches to counterbalance the muscle prolapse.
Fig. 6 Three-Layer Reconstruction With
Iliotibial Tract After Endoscopic Resection Of Sinonasal Tumors: Technical Note Davide Mattavelli Md, Alberto Schreiber, Md; Marco Ferrari, Md; Remo Accorona, Md; Andrea Bolzoni Villaret, Md; Paolo Battaglia; P a o l o C a s t e l n u o v o , P i e r o Nicolai,World Neurosurgery . 2017
Step 2: First Layer
The first layer is oversized by about 30% larger than the dural defect and split anteriorly to adapt to the residual falx cerebri. It is positioned in the subdural space with a blunt instrument (i.e. a 55° Khun-Bolger Curette – Karl Storz).
In this step, there are two possible tricky situations:
1. The presence of brain atrophy (more common, but not limited to the elderly) together with CSF loss can create a large gap between the parenchyma and the ASB; consequently, positioning of the graft can be problematic because it tends to fall behind instead of adhere to the dural layer. In this situation, the use of fibrin glue can be very helpful. It is advisable to fix first the posterior edge and then proceed anteriorly, because a possible redundancy of
the graft is more easily managed at the posterior frontal wall rather than at the planum sphenoidale, in close proximity to optic chiasm.
2. The presence of brain prolapse may facilitate the extrusion of ITT. Even in this situation, fibrin glue is useful to properly fix the graft.
In addition, intradural irrigation with saline solution and antibiotics can replace CSF loss and may help the graft to adhere to dura by restoring the CSF pulsation.
After the first layer is positioned, the onset of CSF pulsation is a simple but effective demonstration of watertight adherence of ITT to the dura.
Step 3: Second Layer
The second layer is inserted in the epidural space. Its size must be just a few mm larger than the defect, so that it can be adequately placed between dura and the residual bony ASB. For this purpose, it is of utmost importance that the epidural space is dissected and that the dura is detached from the bone before dural resection. Otherwise, this maneuver would be far more difficult and ineffective.
The most critical part in positioning the second layer is the posterior edge, because of the slim space for the graft. It is very important to rely on tactile feedback, by exploring with a blunt and thin instrument the space available adherent to the bone and then guiding inside the ITT (i.e. a sharp 15° tip angled spatula– Karl Storz). Otherwise, the risk is to place the graft in the subdural space and dislocate the first layer. Furthermore, ITT must be carefully laid trying to avoid the entrapment of air between the first and the second layer, which could favor extrusion of the grafts.
Of note, a watertight closure of the defect must be achieved with the first two layers; thus, at this moment the reconstruction must be carefully inspected for any sign of CSF leak.
Step 4: Fat Graft
Pieces of fat previously harvested from the thigh are placed on the contour of the bony resection to eliminate any dead space and flatten the residual bony ASB. Moreover, fat is an optimal sealer and promotes graft integration.
Step 5: Third Layer
The third layer is positioned overlay. Its main role is to protect the underlying reconstruction. Care must be taken not to block drainage of the frontal and sphenoid sinuses.
Each layer and the fat grafts can be sealed with fibrin glue. The use of fibrin glue is not strictly necessary, but it may be a valid help in non-ideal situations, which have been previously described.
At the end of the procedure, the frontal sinuses can be stented with rolled silastic sheaths to allow subsequent frontal sinus debridement with no risk for the reconstruction. The surgical cavity is packed with glove fingers filled with Lyofoam (Seton Health Care Group, Oldham, U.K.). 47
6.5 Inlay-Onlay techniques
Published closure techniques, whether using free grafts or vascular flaps, typically encourage the use of a multilayer closure. These approaches often utilize fascia, a biosynthetic dural replacement (e.g., collagen matrix), or fat as a direct subdural (and/or epidural) inlay repair, followed by an onlay repair of free grafting material (free mucosa or alloderm are commonly used) or a pedicled flap extracranially. Mucosal edges and contact points are often reinforced with biologic adhesive and supported with materials such as gelatin foam. 40During the inlay (underlay) technique, the
intact dura is separated from the edge of the skull base defect to expose an adequate buttress for the stabilization of the graft. The free graft, or flap, should be designed in such a way that it can be
pushed a few millimeters between the bone and the dura on all sides of the defect. Bone or cartilage underlay grafts are advocated for large, bony defects associated with herniating brain or meninges. Thus the inlay technique is technically more demanding than the overlay technique. Inlay grafting is suited to repair defects of the posterior wall of the frontal sinus, the ethmoid roof, and, sometimes, the sphenoid sinus. Indeed, most authors reported on the use of free tissue grafts as an "overlay patch" (79% of all cases), as opposed to an "inlay" technique, which was used in only 12% of the cases.
The onlay technique is recommended if there is a risk that nerves or vessels may be damaged when dissecting the dura from the surrounding bone, when inserting the graft, or if an inlay technique is not technically possible. The graft is placed generally over the dural lesion and over the exposed bony margins, which have been denuded of mucosa. The graft is supported in place with layers of Gelfoam/Gelfilm or Surgicel, followed by a packing of gauze impregnated with an antibiotic ointment or some other method of fixation. As an alternative, a vascularized tissue flap may be designed transnasally, using middle turbinate mucoperiosteum or septal mucoperichondrium. 9
An onlay graft that is too small is a common cause for failure and postoperative CSF leak. Failure can likewise result from misplacement of the mucosal surface toward the dural defect, an error that can compromise graft adhesion to surrounding bone. With this in mind, care should also be taken in introducing mucosal tissue into the intracranial space, as this can potentially result in intracranial complications and mucocele formation. Of note, in select cases success has been achieved without covering the entire defect with vascularized tissue, instead using a combination onlay graft in addition to allograft to cover the entire defect . 40
6.6 Gasket seal closure
Gasket seal closure is a method for watertight closure of the cranial base using autologous fascia late held in place by a rigid buttress.
Garcia-Navarro et al 48 described this technique in 46 consecutive patients.
Surgical technique
For the gasket seal to be effective, the defect in the skull base must be surrounded by a rim of bone. The vertical and horizontal diameters of this defect are measured either with a ruler or with a cottonoid. Both fat and fascia lata grafts are harvested from the thigh. If a large intracranial cavity remains after removing the lesion, fat is used to obliterate this dead space to prevent pooling of CSF.
The fascia lata graft is fashioned in the same dimensions of the cranial base defect but with an additional 2 cm of diameter so as to extend 1 cm beyond the edge of the cranial base defect circumferentially. The fascia lata graft is placed over the defect. A piece of vomer, or MEDPOR (Porex Corp, Newnan, Georgia, USA), is cut to be the same size as the defect. This rigid buttress is placed over the fascia lata graft and countersunk into the defect so that the edges of the buttress are wedged just beyond the bony edges of the defect, holding it in place. The center of the fascia lata graft is intracranial, whereas the edges remain in the sinus cavity, similar to a cauliflower leaf (Figure 7).
Fig.7
The fascia lata, which is circumferentially wedged between the bony edge of the cranial defect and the graft creates a watertight gasket seal (Fig. 8)
The philosophy behind the gasket seal was to provide an autologous graft that would be placed in direct contact with the patient’s dura and held firmly in place to facilitate vascularization or fibrosis to ensure a long-term seal. The graft must be held rigidly in place hence the buttress, which avoids the need for an inflated intranasal balloon with its inherent risks of overinflation, local infection, sinusitis, and postoperative distress to the patient. However, if the buttress falls away from the skull base, the gasket seal fails. Likewise, as we learned in one of our failures, if the defect traverses two separate geometric planes, the gasket may fail because the buttress is not curved and exists in only one plane in space. In this situation, use of a sutured graft covered with an NS flap and a balloon for buttressing may be indicated.
Fig.9 Fig.8
Based on the configuration of the defect, which involved both the planum and the clivus and contained a 90-degree angle in the defect, we were unable to attain an adequate gasket seal (Figure 9)48.
6.7 Bath plug technique
The bath plug technique, described by Wormald in 1997, is a method for repairing anterior skull base defects.49 A fat plug, three times the width of the defect, is positioned in underlay in the
epidural or subdural space. It theoretically overcomes the increased ICP by preventing the high pressure from displacing the graft. Sanderson et al.50 stated that the bath plug technique fits and
conforms the shape of the defect, is resistant to infection, and is strongly adherent to bone. However, the bath plug technique is not recommended in defects next to important structures. It leads to difficult interpretation of soft tissue shadows after repair, and its back herniation may prevent osteoneogenesis. 51
Fig.10 Image-Guided Endoscopic Repair of Cerebrospinal Fluid Rhinorrhea by the Bath Plug Grafting Technique Hazem Saleh, ; Sameer Al Bahkaly, Laryngoscope, 2011 51
7. Perioperative managment
Nasal packing is commonly advocated to support the graft in place. Gelfoam or Gelfilm are frequently used to separate the graft from the packing material, to prevent avulsion of the graft or flap during its removal. Whether to use and how long to keep the postoperative packing are based on the surgeon's experience, although most authors recommend removing the packing 3 to 5 days after the surgery.
7.1 Fibrin glue
The efficacy of fibrin glue in preventing CSF leaks remains controversial. Although histopathological studies suggest that fibrin glue may trigger an inflammatory response that may promote healing, studies have reported a success rate of 97% with fibrin glue and 92–100% without glue. 52 Rodney et al., suggested that if tissue adhesives are used, they must be applied
conservatively because a thick layer of adhesive may prevent the graft material from coming in contact with the wound bed. 53
7.2 Lumbar drain
The use of a lumbar drain intraoperatively can aid in graft placement with well timed removal of 10 – 20 ml of CSF. This removal can provide brief decompression of the dura to allow elevation from the bony defect and enable placement of the bone graft in an underlay fashion. Hegazy 9 recently advocated the use of perioperative and post-operative drain placement in situations in which elevated intracranial pressure contributes to leak pathogenesis, including spontaneous and traumatic leaks, as well as leaks with encephaloceles. Many authors recommend the use of postoperative lumbar drain as a means of controlling intracranial pressure and decreasing pressure on the newly placed leak site for 24 – 48 h. 5
The use of lumbar drainage is still controversial. In fact, its invasive procedure, may produces headaches, nausea, meningitis, or pneumoencephalus .54
In cases of spontaneous CSF leak with elevated ICP, a diuretic is often indicated. Schlosser et al 6
reported a mean decrease in ICP of 9.9 cm water approximately 3 to 4 hours after the administration of intravenous (IV) acetazolamide. Patients with an adequate response to IV diuretics are routinely
prescribed long-term oral diuretic therapy, and our previous series of spontaneous CSF leaks emphasized the need for strict management of underlying intracranial hypertension.55 Further
investigation is needed to determine whether chronic use of oral diuretics is effective in maintaining ICP within the normal range, thus preventing CSF leak recurrence.
Although a combination of medical management, lumbar drain, and endoscopic repair of the CSF leak is often successful, cases of refractory elevated ICP and recurrent leakage may require more permanent diversion procedures, such as VP shunts.
VP shunts consist of a ventricular catheter, a shunt valve, and a peritoneal catheter. These shunts are associated with a significant failure rate secondary to complications, including infection, obstruction, mechanical malfunction, and over- or under-drainage.
Patients with significantly elevated ICP, or those who failed to respond to IV acetazolamide, were considered for the VP shunt procedure. 1
7.3 Antiobiotcs
Perioperative prophylactic antibiotics contribute to the low incidence of meningitis following skull base surgery and their use during the repair of CSF leak, regardless of the approach, is indicated.9
8. Success rate
Initial success rates of CSF leak repair ranging from 85 to 90 percent and overall success rates of 97 to 99 percent have been well documented in the literature and demonstrate the efficacy of endoscopic repair. 1
In a meta-analysis of 55 studies involving 1,778 fistula repairs, Psaltis et al., observed a success rate of 90.6% following first endoscopic repair for CSF rhinorrhea, which improved to 96.6% following a second endoscopic procedure.24 The success rate in the largest series of endoscopic repair of CSF
leaks reported by Kirtane et al., was 96.63% following first surgery and 98.88% after revision surgery. 56
The long-term success of endoscopic repair is influenced by CSF leak etiology. Studies indicate increased failure among spontaneous CSF leak repair, especially within the subset of patients demonstrating evidence of increased ICP. 22 In addition to intracranial hypertension, commonly
cited patient factors associated with recurrent CSF leak include obesity, noncompliance, comorbidities, and poor wound healing.
In a review of long-term outcomes of endoscopic repair, Zuckerman et al 57 focused on the timing
of recurrent CSF leaks. The average time for recurrence in their series was 7 months, ranging from 1 to 25 months. Banks et al., observed spontaneous leaks recurring at 7 months (median range: 4 days–24 months) and traumatic leaks recurring at 4 months (median range: 4 days–29 months). 1
9. Complications
Nasal crusting is the most common morbidity and occurs in > 95% of patients (observed at 1 month postoperatively). Patients with severe crusting need aggressive detriment and irrigations, requiring frequent visit to the office. Nonetheless, 50% of patients achieved a crust-free nasal cavity by 3 months postoperatively. 58
Addition sinonasal complications include nasal synechiae (9%) , alar sill burns (5%), maxillary nerve hypesthesia (2%), palatal hypoestesia ( 7%), incisisor hypoestesia ( 11%), serous otitis media ( 2%), taste disturbance (7%) and malodor. Most of these complications are temporary, and most patients recover full nasal function by 6 months after surgery .59
As previously mentioned, resolution of these surgical sequelae requires intense postoperative care. Meningitis, chronic headache, pneumocephalus, intracranial hematomas, frontal lobe abscess, and anosmia can all be a consequence of transcranial or transendoscopic repair. 1
10. Objective of the review
Cerebrospinal fistula might occur spontaneously or as a results of a head trauma, surgery injury or of a different pathological conditions such as head neoplasm, inflammatory diseases eroding the skull base or congenital malformations. Over the past fifty years, CSF closure techniques have undergone several and significant evolution and since 2000 advancements in endoscopic technology have led to consolidation of the trans-nasal endoscopic approach.
Despite multiple publications site low incidence of complications and low rates of recurrence in association with this minimal invasive techniques, meaningful informations about the ideal method or material to use and about the timing for repair, still lack in literature.
The purpose of this review was to summarize the success rate for endoscopic CSF leak repair as well as whether specific techniques or materials influence the primary success rate. In addition, the impact of epidemiology, etiology, clinical presentation, site and dimension of the fistula, management success-related factors, including timing to perform surgery and information on the effect of adjunctive measures such as lumbar drain and intrathecal fluorescein, were analysed through a review of the latest advances in endoscopic CSF management published in the past 10 years.
11. Materials and methods
The systematic review was performed using independently developed search strategies in literature review methodology, and it was written in accordance with PRISMA Statement, to guarantee a scientific strategy of research to limit bias by a systematic assembly, critical appraisal and synthesis of all the most relevant studies published on this topic 60,61.
The databases interrogated included PubMed Clinical Queries http://www.ncbi.nlm.nih.gov. Reference lists from identified articles were searched and cross-referenced to identify additional relevant articles, and national experts in the field were contacted to identify unpublished data.
The search terms included the following various combinations to maximise the yield: Anterior skull base defect AND Anterior cerebrospinal fluid leaks AND Endoscopic reconstruction AND Sinonasal malignancies AND Reconstructive technique AND Double flap technique AND Nasoseptal flap AND Inferior turbinate flap AND Medial turbinate flap AND Multilayer reconstruction AND Iliotibial graft AND Graft.
The search was performed for the first time on December 2017 and was set to automatically update periodically until July 2018.
First, duplicates were removed electronically. Then abstracts were reviewed to exclude obviously irrelevant articles. Non-English language papers, experimental studies and cases report were excluded.
The inclusion criteria were set a priori and deliberately kept wide to encompass as many articles as possible without compromising the validity of the results, and they included articles: (1) published from 2007 onwards; (2) reporting published series of > 10 patients underwent endoscopic repair of CSF leaks; (3) about endoscopic trans-nasal cerebrospinal fluid leaks repair; (4) excluding open approaches; (5) reporting data distinguishes results of endoscopic procedures divided into: multilayer or single layer; (6) considering different endoscopic technique: NSF, ITF, MTF, Fascia Lata, Abdominal fat graft; (7) with a clear description of CSF leaks location (ethmoid, sphenoid, frontal and multiple site); (8) dividing CSF leaks etiologies : iatrogenic, neoplastic ( benign tumors, malignant tumors) and spontaneous; (9) excluding traumatic etiology.
We filtered the studies to ensure that only data from centers that had published on at least 10 patients were included in the review; this was done as a quality assurance measure as there are several case series in the literature, which have published the results of small numbers of cases spanning several years.
We chose as success rate of endoscopic CSF leaks repair at first surgery as the primary measure of oncological outcome.
Abstracts were analyzed to identify papers that fulfilled inclusion criteria and a first qualitative and descriptive review-analysis of selected articles was carried on; whilst, exclusively, publications clearly describing their aim and objectives, their inclusion and exclusion criteria, with clear or detachable statistical data, reporting success rates and well describing the surgical techniques and post-operative complications, were included in our meta- analysis.
11.1 Study characteristics and quality assessment
All included papers were graded using the NICE scoring scale for retrospective case series (Available at: http://www. nice.org.uk/nicemedia/pdf/Appendix_04_qualityofcase_
series_form_preop.pdf). This is a scoring scale with eight items, with each item scoring zero or one
based on the study methods (Yes = 1; No = 0). Scores of ≥ 6 are considered to indicate a good quality study, scores between four and five as fair and those studies with a score of three are treated as poor quality (Table 3).
11.2 Statistical analysis
Fisher’s exact test was used for statistical analysis of categorical data for the descriptive review, and a value of p < 0.05 was considered significant.
The pooled estimate of each statistic was calculated after Freeman–Tukey double arcsine transformation to stabilise the variances. A random effect model was specified, using the method of DerSimonian and Laird, with the estimate of heterogeneity being taken from the inverse-variance fixed- effect model. Heterogeneity is also quantified using the I-squared measure.
All analyses were performed using STATA version 13 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP).
The possibility of publication bias was evaluated with the Begg and Egger tests, as well as visual inspection of the funnel plot. When possible, results are described in accordance with the PRISMA guidelines for reporting meta-analyses, with 95% confidence intervals reported throughout.
When studies have low heterogeneity (pragmatically, I < 25%), the differences between reported outcomes can be explained simply by the observed natural differences between patients. In this case, we can consider that all patients are part of the same larger pool. A fixed-effects meta-analysis is appropriate in which each patient is given approximately equal weight.
12. Results:
The various stages of systematically assessing the abstracts and reasons for exclusion from the review are described in Fig. 11:
Fig.11
Statistical assessment was performed primarily with descriptive data. Selected studies were summarized in Table 4:
As for their quality, as assessed by the quality assessment criteria outlined above, 4 papers scored eight or seven, 9 scored six, 5 scored five and 10 papers scored four or three. Of this 5 were prospective studies and 23 retrospective studies or review. The study periods range from 12 to 252 months with a median of 72 months.
12.1 Demographics
A total of 1767 patients were identified, of this 937 were male and 857 female (Fig. 12). The median age was 51 years old (range 15-86 years).
Fig.12
12.2 Factors relating to CSF leak
• Size
The defect size was reported in 11 papers, the median defect size was 2,81 cm2 . In 395 of 667
patients (59%) the defect size was < 2cm , in the remain 272 patients (41%) the size was > 2cm (Fig. 13).
Fig.13 • Etiology
Etiologically, CSF leaks was divided into 3 categories : neoplastic (1187 patients 67%), spontaneous (223 patients 14%) and iatrogenic (341 patients 19%) CSF leaks. (Fig.14)
Fig.14
Because of individual data were not available, we have looked at the macro-level correlations between articles. B en ig n Tu m or s Ia tr og en ic C S F M al ig na nt T um or s S po nt an eo us C S F F re qu en cy
Pareto Chart for Etiology
0 50 0 10 00 15 00 80 % C u m u la tiv e P er ce n ta ge < 2cm > 2cm
Our data have shown that there was a relationship between etiology of CSF leaks and its size > 2 cm. Table 5 shows that there was a statistically significant difference in the group of neoplastic and spontaneous CSF leaks compared by the dimension of CSF leaks (Fig. 15-16).
Table 5
ETIOLOGY coefficient p value
Neoplastic -0.58 0.25*
Iatrogenic -0.19 0.88
Spontaneous -0.98 0.01*
Fig.15
• Location Site
In 488 (30%) cases the leak was located in the ethmoid roof, in 956 (59%) patients in sphenoid sinus , for 70 (4%) patients in frontal sinus. Multiple location of CSF leaks was found in 115 cases (7%).(Fig.17)
Fig.17
Table 6 indicate the ethmoid location as the predictive factor more closely associated with CSF leaks size >2 cm. (Fig.18)
Table 6
LOCATION SITE Coefficient p value
Ethmoid 0.85 0,03*
Sphneoid 0.74 0.09
Fig.18
12.3 Factors relating to adjuvant management
• Intrathecal fuorescinFluorescin injection for detection of CSF leaks was used in 589 patients (33,3%)
• Lumbar drain
16 studies provided complete data on the use of lumbar drains. In these studies 703 lumbar drains were used. The duration of the use varied from 1 to 8 days, with most studies reporting use between 2 and 5 days. The benefit of lumbar drain usage could not be calculated from the limited data provided by the studies.
12.4 Factors relating to repair
A multilayer reconstruction was fashion in 1405 patients , in the other 394 was used only single layer. (Fig.19)
Fig.19
Various techniques were adopted in single layer or as a part of a multilayer reconstruction (Fig20). In our series fascia lata was performed in 1085 patients (61%), NSF in 979 (55%), Janus flap (bilateral nasoseptal flap) in 20 (1,1%).
MTF was harvested in 14 patients (0,7%), ITF in only 7 cases (0,4%).
Abdominal fat graft was carry out in 836 patients (47%) instead HVPMMA (High viscoity polymthylmethacrylate), a variety of injectable cements, was used in 12 patients (0,7%).
Fig.20
Multilayer Single Layer
Fascia Lata NSF Janus Flap MTF ITF Fat Graft HVPMMA
• Success rate
28 studies were identified with complete data available on the primary success rate for endoscopic repairs. Pooling data from 1767 primary repairs, the success rate was 96% .
CSF leak recurrences were found in 71 (4%) patients.
Median time of recurrence was 66 days (range 2-1095 days). Revision surgery was required in 55 cases (3%).
Heterogeneity of the results between the studies is assessed graphically by forest plots and statistically using the quantity I2 that describes the percentage of total variation across studies that is attributable to heterogeneity rather than chance (Fig.21).
Fig.21 RE Model 0.6 0.7 0.8 0.9 1 Observed Outcome Martinez−Capoccioni et al. Fyrmpas et al. El−Banhawy et al. Nyquist et al. Luginbuhl et al. Germani et al. Tabaee et al. Eloy et al..3 Moliterno et al. Banks et al. Eloy et al..2 Garcia−Navarro et al. Sciarretta et al. Hoffmann et al. Nix et al. Gilat et al. Sprekelsenet al. Kassam et al. Regi et al. Ramakrishnam Mattavelli et al. McCoul et al. Eloy et al..1 Sannareddy et al. Gruss et al. Fonmarty et al. Patel et al. El−Sayed et al. 0.97 [0.92, 1.03] 0.91 [0.74, 1.08] 0.91 [0.83, 0.99] 0.94 [0.85, 1.02] 0.90 [0.77, 1.03] 0.97 [0.92, 1.02] 0.91 [0.86, 0.96] 1.00 [1.00, 1.00] 1.00 [1.00, 1.00] 0.91 [0.87, 0.95] 0.99 [0.96, 1.01] 0.96 [0.90, 1.02] 0.92 [0.87, 0.97] 0.96 [0.89, 1.04] 0.86 [0.78, 0.94] 0.83 [0.68, 0.98] 0.95 [0.91, 0.99] 0.89 [0.82, 0.96] 0.91 [0.83, 0.98] 1.00 [1.00, 1.00] 0.94 [0.91, 0.98] 0.97 [0.93, 1.00] 1.00 [1.00, 1.00] 0.81 [0.69, 0.93] 0.96 [0.91, 1.01] 0.96 [0.90, 1.03] 0.96 [0.93, 0.99] 1.00 [1.00, 1.00] 0.96 [0.94, 0.97]
As shown below, high figures indicate greater homogeneity index in the data (Fig.22): Fig.22
Both in the forest plot and in the funnel plot (Fig.22), standard errors of the success rates are estimated through the classic estimator of the standard error of a proportion.
From statistical analysis, significant predictive factors were not been found regarding the onset of CSF leak recurrence (Table 7):
Funnel plot − Success rate
Observed Outcome Standard Error 0.087 0.065 0.043 0.022 0 0.8 0.9 1 1.1
Table 7
12.5 Factors relating to post-operative period
Post operative nasal packing was used in 1207 patients (69%).
Median hospital stay was of 8 days (range 5-11 days).
We used a linear probability model taking length of hospital stays longer than 7.5 days as a response variable. Results show that “ethmoid location”, “sfenoid location” and “malignant tumors” are predicting factor for the length of the hospital stay (Table 8):
Table 8
Variable Coefficient pvalue
intercept -0.09 0.45 size >2cm -0.05 0.52 ethmoid location 0.14 0.42 sfenoid location 0.10 0.44 benign tumors 0.04 0.22 malignant tumors 0.11 0.68
prior radiation therapy 0.00 0.91
Variable Coefficient pvalue
intercept 0.90 0.01 *
ethmoid location -1.06 0.02 *
sfenoid location -0.73 0.04 *
benign tumors -0.11 0.66
malignant tumors 0.68 0.06 °
Complications (Fig.23) occur in a total of 81 patients (4,5%), of these 43 were categorized in major complications (meningitis, pneumoencefalo, hemorrhage, intracranial hypertension), 38 in minor complications (headaches, post operative sinusitis, crusting, sinechiae, hematoma, mucocele…).
Fig.23
!
Response variable: proportion of individuals who had major complications.
All the variables refer to the proportion of individuals with that characteristics over the total of individual that took part to the study in each article.
All the factors considered are somehow related to the proportion of individuals with major complications. The strongest predictors are tumors and prior radiation therapy (Table 9).
Since we are not using individual-level data, we cannot draw individual-level conclusions, but we can only conclude that, on average, studies with higher numbers of patients with a benign tumor, also reported higher numbers of patients with major complications.
M e n in g iti s H a rv e st in g o r su rg e ry − re la te d D o n o r si te r e la te d P n e u m o e n ce fa lo H e m o rr h a g e S ys te m ic − c o m o rb id ity In tr a cr a n ia l H yp e rt e n si o n F re q u e n cy
Pareto Chart for Complication
0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 % C u m u la ti ve P e rc e n ta g e
Table 9
Conversely, significant predictive factors have not been found regarding the onset of minor complications (Table 10):
Table 10
Variable Coefficient pvalue
intercept 0.11 0.04 * size >2cm 0.06 0.05 ° ethmoid location -0.14 0.06 ° sfenoid location 0.15 0.02 * benign tumors 0.13 <0.01 *** malignant tumors -0.30 0.01 **
prior radiation therapy -0.01 <0.01 ***
Variable Coefficient pvalue
intercept -0.09 0.45 size >2cm -0.05 0.52 ethmoid location 0.14 0.42 sfenoid location 0.10 0.44 benign tumors 0.04 0.22 malignant tumors 0.11 0.68
13. Discussion
Cerebrospinal fluid leak denotes a skull base fistula connecting the subarachnoid space to the nasal cavity.
Irrespectively of their iatrogenic, traumatic, spontaneous or tumorous nature, skull base defects have to be closed meticulously.
Over the past 20 years more invasive open approaches have been largely replaced by entirely endoscopic ones that have revolutionized the surgical management of sinonasal and skull base diseases. Improved cosmesis, olfaction preservation, and reduction in morbidity are among a few of the cited reasons for the preference of endoscopic repairs.
The incidence of post-operative CSF leak is generally considered to be the most relevant primary outcome measure reflecting the success of skull base reconstruction62.
The aim of this study was to ascertain the outcome after transnasal endoscopic repair of CSF leaks and to identify factors regarding patient, CSF fistula, and different endoscopic surgical techniques that may influence the results of the repair.
Several studies have therefore analyzed the most important risk factors associated with skull base reconstruction failure, identifying as possible factors involved the size of the defect63, the
anatomical site of the defect, the presence and flow rate of intraoperative CSFL 64 , the surgical
technique used (free grafts versus vascularized flaps) 65, the impact of previous or subsequent
radiotherapy 65and the absence of a bony rim around the defect.
Independent analysis of these factors individually was difficult, primarily because of the potential interdependence of the criteria, with certain locations of the skull base more characteristically associated with neoplastic, spontaneous or iatrogenic etiology.
When we compared reconstruction outcomes by subsite of defect, there were no clear differences between vascularized versus non vascularized reconstruction techniques for any individual subsite.
It appears that specific characteristics of the defect may be more relevant of selection of reconstructive techniques, rather than anatomic subsite.
For example, in anterior cranial base defects, it is thought that the weight of the brain on an inlay graft helps to hold the material in position securely and prevent migration. In contrast, the clivus is anatomically closer to the anterior brain cisterns and ventricles. Resection of lesions in these areas is therefore at higher risk of resulting in high-flow CSF leaks, and hence, the lesions may benefit more from vascularized flap reconstruction.
A retrospective study published by Gruss et al., revealed that the dimensions of the dural defect correlates with recurrences of CSF leak 63. Flap surgery is highly successful with small defects with
a flap failure rate of 3.8% for defects less than 2.0 cm2. However, 16.7% of defects 2.0 cm2 or larger
failed (P=.031).
In a systematic review of the literature, Harvey et al. concluded that for skull base defects larger than 3 cm reconstruction with a vascularized flap had a significant advantage over free grafting in preventing post-operative CSFL66.
Small defects (<10 mm), instead, generally do not need repair with vascular flaps 67.
Data emerging from a study performed by Turri-Zanoni et al. 62 outline that it is not the defect size
that makes the reconstruction complex. Much more important is whether the borders of the defects can be identified and exposed, since this is what determines the ease, complexity or even impossibility of the procedure.
Thus, a huge ASB defect extending from orbit to orbit and from the planum sphenoidalis to the frontal sinuses may be easier to repair than a much smaller defect in which the precise borders are hardly identifiable.
The absence of these borders precludes the positioning of the inlay grafts, namely the first intracranial intradural and the second intracranial extradural (in a sort of “epidural pocket”), and so significantly increases the risk of post-operative leakage.
