2-9 Reda Kamel

7/28/2019 2-9 Reda Kamel

1/8

Cerebrospinal fluid rhinorrhea: classification and guidelines for endoscopic repair

2 PAN Arab Journal of Rhinology

PAN Arab Journal of Rhinology, Vol. 1, No. 1, October, 2011

Cerebrospinal fluid rhinorrhea: classification and guidelines forendoscopic repair

Reda H . Kamel,1Elgohary M . Elgohary,2Tarek M . Kandil ,3Khal ed S. Anbar,2M ahmoud A. M okbel,1Hatem B.,1

Hany A . Al gamal41Otorhinolaryngology Department, 2Neurosurgery Department, Cairo University, 3Cairo University Student's Hospital,4Al-Monira Hospital, Egypt

Correspondence to: Reda H. Kamel, Email: [email protected]

Background: Cerebrospinal fl uid (CSF) rhi norrhea may be tr aumati c, developmental , pathol ogical, or spontaneous.D i fferent rout es were suggested to approach the defect and many t echniques w ere int roduced for repair .

Objective: To present a new classif i cati on of CSF rhi norr hea based on the det ai l ed sit e of t he skul l base defect and t odemonstrat e how to ut i l ize it i n determi ning the best approach for repair .

Methods: N i net y f our cases of CSF leak hav i ng 103 skul l base defect s were classif i ed accordi ng to t hei r detai l ed sit es. Theapproach fol l ow ed t o repair t he defect w as designed accordi ng t o the si t e of t he l eak. These approaches i ncluded the

endoscopic tr anscri bri form , tr anset hmoi d, axil lar y f l ap technique, transnasal t ranssphenoid, tr anspterygopalat i ne, direct

tr ansnasal and external ost eoplasti c flap.

Results: Transnasal endoscopi c approach w as useful to approach al l sit es except the l ateral front al sinus defect s, w hereexternal osteoplasti c fl ap was mandatory . Transcri bri form approach w as effect iv e i n anteri or and posteri or cribri form

plat e defects. Transet hmoi d approach w as useful i n anterior and posteri or ethmoids l eaks. Axi ll ary f l ap technique w as

suffici ent i n medial front al sinus and front al recess l esions. Transnasal tr anssphenoid approach w as effi cient i n centr alsphenoid sinus defects and t ransptery gopalat i ne fossa i n l ateral recess leaks. Di rect tr ansnasal approach was resorted to

i n case of absent mi ddl e t urbi nat e. These di ff erent appr oaches offered enough exposure of the defect for r epai r. Pri mar y

closure w as achieved i n 93.9% of defect s and secondary closure i n 100%.

Conclusion: Classifi cati on of CSF rhi norrhea according t o t he det ail ed sit e of skull base defect helps sel ect t he most di rectand least destructi ve approach wi th effecti ve repair .

Keywords: Cerebrospinal fl ui d rhi norr hea, CSF leak, skul l base defect , repai r, classif i cati on, endoscopi c surgery .

INTRODUCTION

Cerebrospinal fluid (CSF) rhinorrhea may be traumatic,

pathological, developmental or spontaneous.(1,2) The leakmay be located at the ethmoid roof, cribriform plate,posterior wall of frontal sinus, or the sphenoid sinus.(2,3,4)

Surgical repair may be achieved transcranially through abifrontal craniotomy, extracranially through an externalethmoidectomy or osteoplastic flap, or transnasally withmicroscopic or endoscopic visualization.(5)

In recent years the popularity of endoscopic closure ofCSF leak has continually increased and in most cases,endoscopic nasal surgery has almost completely replaced

more traumatic transcranial and extracranialprocedures.(3,6,7)

The endoscopic technique offers a direct access, exactidentification of the site of the dural tear and preciseplacement of the graft. It permits preservation of thefunctional anatomy of the nose including smell.8,9 It

7/28/2019 2-9 Reda Kamel

2/8

Reda H. Kamel, et al.

PAJR, Vol . 1, No. 1, Oct ober, 2011 3

offers shortened operating time in conjunction withsuccess rates of 90% after primary attempts and 97% aftersecondary repair.(3,10,11)

The most commonly used classifications of cerebrospinalfluid rhinorrhea are based on the etiology.(1) There is noclassification based on the detailed site of skull basedefects. Moreover there are no clear guidelines for theoptimum approach utilized in each defective site.

The aim of work of this study is to present a newclassification of skull base defects based on the detailedsite of the leak and to demonstrate how to utilize it as aroadmap to select the most appropriate approach for leakrepair.

PATIENTS AND METHODS

This study included cases of CSF leak presenting to theENT outpatient clinic of Cairo University Hospitalbetween July 1997 and August 2011, operated upon by the

first author. Cases with persistent leak for more than 3months refractory to conservative therapy were included.Cases of iatrogenic leak or leak associated withmeningoceles or meningoencephalocele were alsoincluded.

Cases of comminuted fractures of skull base, cases withprior attempts for repair, and patients with CSFrhinorrhea of temporal bone were excluded.

Endoscopy was performed in all cases before surgery toassess the nasal cavity, follow the leak to the affected site,and to detect any associated pathology. Immunofixationof beta-2-transferrin was resorted to in suspicious cases.To define the site of the leak preoperatively, highresolution CT and/or CT cisternography were performed.In case of suspected meningocele ormeningoencephalocele, MRI study was performed.

Ninety four cases of CSF leak were operated upon undergeneral anesthesia. In case of defects of the cribriformplate, the leak was approached via the transcribriformapproach, medial to the middle turbinate in case ofanterior cribriform or medial to the superior turbinate incase of posterior cribriform. In case of defects of the foveaethmoidalis or lateral lamella, the transethmoid approachwas utilized where the uncinate process was excised andthe bulla ethmoidalis was removed to expose defect. Incase of skull base defect of posterior ethmoid, thetransethmoid approach was also utilized. After anteriorethmoidectomy, the basal lamella was penetrated and theposterior ethmoids were cleared to expose roof. In case ofdefects of the frontal recess or medial part of the posteriorwall of the frontal sinus, medial to the sagittal planethrough lamina papyracea, axillary flap was performedwhere the uncinate process was excised and agger nasi

cells were removed with preservation of the bullaethmoidalis. In case of defects of the central sphenoidsinus either from the planum sphenoidale, sellar floor orclivus, the transnasal transsphenoid approach wasfollowed. The sphenoid sinus ostium was identifiedbetween the superior turbinate and nasal septum andwidened. In case of defects of the lateral recess of thesphenoid sinus, the transpterygopalatine fossa approachwas followed. Anterior and posterior ethmoidectomy,wide middle meatal antrostomy and sphenoidmarsupialization were performed. The posterior wall ofthe maxillary sinus and anterior lip of the sphenopalatineforamen were drilled. The periosteum of thepterygomaxillary fossa was then displaced laterally toexpose the vidian nerve and canal then the maxillarynerve lateral and superior.

In case of defects of the lateral part of posterior wall of thefrontal sinus, lateral to the sagittal plane through thelamina papyracea, external osteoplastic approach wasperformed. (Fig. 1) shows the different detailed sites ofCSF leaks and approaches used for repair. In case of

absence of middle turbinate (4 cases, 4.25%), due todisease or prior surgery, direct transnasal approach wasadopted.

The site of the defect was confirmed intra-operatively, byidentification of the leak and washout sign. Blue-lightfilter and pre-operative intrathecal fluorescein wereutilized in one case. In case of failure to define the site ofthe leak, intraoperatively, Valsalva-like maneuver, andintra-operative intrathecal fluorescein [2 cases] wereperformed.

After identification of the defect, the surrounding bonewas exposed by mucosal dissection. In case of cribriform

defects, due to the tiny space available, the mucosaaround the defect was cauterized to expose the bone. Insmall defects 5 mm, nasal septal mucoperichondriumgraft was harvested from the contra lateral side. In case oflarge defects > 5mm, fascia lata was utilized. Fat wasadded in lateral frontal sinus, sphenoid sinus and largefovea ethmoidalis defects. Graft was applied in an on-layfashion and Gelfoam was put on top. Sofartule was thenintroduced followed by anterior nasal pack. The anteriornasal pack was left for 2 days and the sofratule for 5 days.All patients received peri-operative antibiotics and nolumbar drain was used.

In case of meningocele or meningoencephalocele, it was

cauterized using bipolar cautery to ablate prolapsed tissuetill the site of the defect was identified and repaired.

Image guided system [IGS] was utilized in leaks of thelateral recess of sphenoid sinus.

7/28/2019 2-9 Reda Kamel

3/8



I

II

III

IV

III

VVI

2 1

3

6

4

5

7 8

11

9

1210

Fig. 1 show s the di fferent det ai l ed sit es of CSF

rhi norrhea and t he approaches uti l ized for r epair . Sit es

of skul l base defects: 1, 2 posteri or w al l of t he front al

sinus, l ateral and medial t o the sagi t tal plane thr ough

lamina papyracea respectively, 3. frontal recess, 4.

fovea ethmoidalis, 5. lateral lamella, 6. anterior

cribriform plate, 7. posterior cribriform plate, 8.

posterior ethmoid, 9. planum sphenoidale, 10. sellar

fl oor, 11. cl i vus, 12. l at eral recess. Appr oaches used i n

the study: I. external osteoplastic fl ap, II . axill ary fl ap,

III . Transethmoid [anterior and posterior], IV.

transcribriform, V. transnasal transsphenoid, VI.

t ranspt ery gopalat i ne fossa. I n case of absence of

mi ddle tur binat e due t o di sease or pri or surgery (2

cases, 3.8%), t he di rect t ransnasal appr oach was

adopted.

RESULTS

Ninety four cases of CSF leak were included. There were30 males (32%) and 64 females (68%). Age ranged between6 and 72 years [mean 41 years]. Eighty nine casespresented by unilateral rhinorrhea (94.7%) and 5 bybilateral rhinorrhea (5.3%). Unilateral rhinorrhea was

from the right side in 52 cases (55%) and left in 37 cases(39.4%). Forty four cases (46.8%) had a history ofmeningitis. The period of CSF leak prior to surgery rangedbetween 3 months and 7 years [average 15 months].

Table 1 shows the etiology of CSF leak. Table 2 shows thesites involved by the skull base defect in general and indetails. It was noticed that the most common site involvedwas the anterior ethmoid (40%) followed by cribriformplate (30%). Concerning the detailed site, the mostcommon site was the lateral lamella of ethmoid bone(26%) followed by anterior cribriform plate (24%) andthen fovea ethmoidalis (14%).

Table 1 shows the etiology of CSF leak in the 94 casesof skull base defects.

Aetiology No of patients %

SpontaneousDevelopmentalAccidental traumaticIatrogenicPathological

472014103

5021

14.910.61.2

Total 94 100

Table 2. Shows the general and detailed sites of skull base defect in the 94 cases of skull base defects.

General site of skull base defect No % Detailed Site of defect No %

Frontal sinus 11 9.6 Lateral 4 3.5Medial 2 1.8

Frontal recess 5 4.4

Anterior ethmoid 46 40 Fovea ethmoidalis 16 14Lateral lamella 30 26

Cribriform plate 34 30 Anterior 27 24Posterior 7 6.1

Posterior ethmoid 9 7.9 Posterior ethmoid 9 7.9

Sphenoid sinus 14 12 Planum sphenoidale 4 3.5Sellar floor 3 2.6

Clivus 2 1.8Lateral recess of sphenoid sinus 5 4.4

Total 114 100 114 100

7/28/2019 2-9 Reda Kamel

4/8



Bilateral defects were encountered in 5 cases (5.3%). It wasnoticed that the lateral lamella was also the most commonsite affected. Bilateral cases had separate defects in fourcases and in the fifth; a large meningoencephalocele washerniating through one wide defect affecting the skull basebilaterally (Fig. 2).

Table 3 shows the number of defects, sites, operations and

procedures performed. Table 4 shows the approachfollowed in each defective site. The size of the defect wasprecisely defined intra-operatively. It ranged between 2 x2 mm and 20 X 25 mm [average 4 x 6 mm]. Fifty cases(53.2%) were associated with meningocele.

Four cases with follow-up less than 3 months wereexcluded. Follow up ranged between 3 months and 14

years [average 52 months]. Six out of the 90 cases (6.6 %)showed persistence (1 case) or recurrence (5 cases) of leakwithin 3 months. Primary successful closure was achievedin 93 out of the 99 defects (93.9%). The sites of failures inthe recurrent cases were medial frontal sinus, lateralfrontal sinus, Fovea ethmoidalis and planum sphenoidale.The case of medial frontal sinus defect was re-operatedupon successfully via the axillary flap with a follow-up of68 months (Fig. 3). The other four cases refused anyfurther interference. The case of persistent leak was thecase of failure to identify the associated ipsilateral defectat the anterior cribriform plate, which was successfullyrepaired one month after primary surgery. Success wasachieved in 100% of cases after secondary repair.

Table 3. Shows the number of defects, sites, operations and procedures performed.

Patients Defects Sites Operations Procedures

Unilateral defect:

Single defect:o One siteo Two siteso Three sites

Double defectsBilateral defects:

Separate defects Continuous defect

89

5

84

5

41

7923

7923

10

81

7949

10

84

79237

41

79239

81

Total 94 103 114 96 102

Table 4. Shows the detailed approach [102 procedures] followed in each defective site. Trans-PPF=transpterygopalatine fossa.

Site of Skull Base Defect Approach followed No %

Lateral frontal sinus

Medial frontal sinusFrontal recess

Anterior cribriform platePosterior cribriform plate

Fovea ethmoidalis of anterior ethmoidLateral lamellaFovea ethmoidalis of posterior ethmoid

Planum sphenoidale

Sellar floorClivus

Lateral recess of sphenoid sinus

Multiple sites & missing middle turbinate

External osteoplastic flap

Axillary flap

Transcribriform

Transethmoid

Transnasal transsphenoid

Trans-PPF

Direct transnasal

Total

4

7

27

46

9

4

5

102

3.9

6.9

26.5

45.1

8.8

3.9

4.9

100

7/28/2019 2-9 Reda Kamel

5/8



Fig 2. Show s t he CT of a case of l arge meningocel e

inv olvi ng the anteri or cribri form pl ate on both sides and

t he lat eral l amel l a and fovea et hmoi dali s on the l eft side.

I t w as approached vi a a left di rect t ransnasal approach.

Fig 3. Show s the CT of a case of t he medi al part of t he

posterior w all of t he frontal sinus that could be handled

vi a axil lary fl ap approach.

DISCUSSION

Classifications are needed to plan surgery and compareresults. CSF rhinorrhea cases are classified according totheir etiology, size or site.(1,12,13-15) In 1968, Ommaya et al.,classified CSF rhinorrhea into traumatic and non-

traumatic etiologies.(1) The traumatic category wassubdivided into accidental and iatrogenic. In thenontraumatic category a distinction was made betweenhigh pressure leaks, as may be due to tumors orhydrocephalus, and normal pressure leaks, with etiologiesincluding osteomyelitic erosion and congenital anomalies.According to the size of the dural defect, they weredivided into small (2.0cm).12 Depending on the site of the leak, skullbase defects were divided into, frontal, ethmoidal,sphenoidal and olfactory cleft defects.(13-15)

The authors present a new classification of CSF rhinorrheabased on the detailed site of skull base defects (Fig. 4).

CSF leak may be anterior or posterior to the basal lamellaof middle turbinate.

Anterior leak is medial or lateral to the vertical part ofmiddle turbinate. Medial leak is from the anteriorcribriform plate. Lateral leak is anterior or posterior to theanterior wall of bulla ethmoidalis. Posterior leak is fromthe lateral lamella or fovea ethmoidalis. Anterior leak ismedial or lateral to the sagittal plane through the laminapapyracea. Anterior medial leak is from the frontal recessor medial part of posterior wall of the frontal sinus.Anterior lateral leak is from the lateral part of theposterior wall of the frontal sinus.

Posterior leak is medial or lateral to the superior turbinate.Lateral leak is from the roof of posterior ethmoid. Medialleak is anterior or posterior to the face of sphenoid i.e.from the posterior cribriform plate or the sphenoid sinusproper respectively. Sphenoid sinus leak is central orlateral. Central sphenoid sinus leak may be from theplanum sphenoidale, sellar floor or clivus. Lateralsphenoid sinus leak is from the lateral recess of sphenoidsinus.

In most prior studies, each sinus was considered as oneunit with no details of leaking sites.(2,5,8,13-14,16-19) Someauthors divided leaking sites into five regions; frontal,anterior ethmoids, posterior ethmoids, cribriform plateand sphenoid with the anterior ethmoid and cribriformmost common.(18-19) Others considered four regions afterthe exclusion of frontal lesions with cribriform mostcommon.(16) Others globally categorized defects into fourregions; frontal, ethmoids, cribriform plate and sphenoidwith anterior ethmoid and sphenoid most common.(13-14)Few considered only three regions after frontal sinusexclusion with ethmoids, sphenoid and cribriform most

7/28/2019 2-9 Reda Kamel

6/8



common in different studies.(2,5,8,17) In the current studyleaking sites were divided regionally into five main sitesand locally into 12 detailed locations. The five regionalsites were frontal, anterior ethmoids, posterior ethmoids,cribriform plate and sphenoid regions. The 12 detailedsites included: lateral and medial parts of posterior wall offrontal sinus, frontal recess, fovea ethmoidalis, laterallamella, anterior cribriform plate, posterior cribriformplate, posterior ethmoid, planum sphenoidale, sellar floor,clivus, and lateral recess of sphenoid sinus. The mostcommon regional site of skull base defect was anteriorethmoids (46.8%) followed by cribriform plate (29%). Themost common detailed local site was the lateral lamella(32.3%) followed by the anterior cribriform plate (25.8%)and fovea ethmoidalis (14.5%).

The choice of the approach depends on the site, size,and/or etiology of the dural defect.(4,7,8,15,17) the currentstudy, the approach followed for repair wastailored according to the preoperative defined detailed siteof leak.

It was suggested that leaks in the cribriform plate andethmoid roof are treated with the standard transnasalendoscopic approach. Complete endoscopicethmoidectomy and maxillary antrostomy are performedfor adequate exposure of the skull base. Frontalsinusotomy, sphenoidotomy and middle/superiorturbinectomies are performed if needed for additionalexposure and to avoid postoperative mucoceles.(14) Othersstated that for defects in the medial lamella of thecribriform plate, it is preferred to go directly between themiddle turbinate and the septum, localize the defect, andseal it without sacrificing the turbinate. For defects in thelateral lamella of the cribriform, anterior ethmoidectomyis done to explore the defect and seal it, also without

sacrificing the middle turbinate.(7,17) In the current study,the anterior and posterior cribriform plate defects wereapproached and repaired directly via transcribriformapproach without any ethmoid workup or insult tomiddle or superior turbinate. Anterior ethmoiddefects were handled via transethmoid approach afterremoval of the bulla, with utmost respect to theintegrity of the middle turbinate. Posterior ethmoiddefects were repaired via transethmoid approach afterremoval of bulla ethmoidalis and penetration of the basallamella without any insult to the superior turbinate.There was no need for sphenoidotomy,maxillary antrostomy or frontal sinusotomy in any ofthese sites.

In central sphenoid sinus CSF leaks the transethmoid ordirect parasagittal approaches were advised.(7,15) Whiledefects located in the lateral recess of the sphenoid sinuswere approached via endoscopic transpterygoidapproach.(20) In the current study the direct transnasal

transsphenoid approach was performed in sphenoid sinuscentral leaks and the transpterygopalatine fossa approachin lateral recess defects.

It was stated that the endoscopic repair of frontal sinusCSF leaks is rarely possible because of reduced access andvisibility. Only small defects of the area that can be seenbulging anteriorly when the frontal recess has beenopened can be dealt with endoscopically.(21) Frontal sinusCSF leaks were divided into 3 anatomic sites: immediatelyadjacent to frontal recess, direct frontal recess and frontalsinus proper.(22) Schlosser and Bolger (2006) advised acombined above-and-below approach in direct frontalrecess defects using endoscopic and open techniques.(15)In the five cases of pure frontal recess defect the axillaryflap technique was performed and the agger cells wereremoved to successfully identify the defect and repair it.The Bulla ethmoidalis was left intact for graft support.Frontal sinus defects located superior or lateral requireosteoplastic flap with or without obliteration.(15) The sameapproach with obliteration was followed in the currentstudy in 4 cases of laterally located frontal sinus CFS leaks.

Endoscopic approach with adjuvant frontal trephinationand/or an endoscopic modified Lothrop procedure wereadvised for unique cases of leaks located in the frontalsinus proper possibly with extension into the frontalrecess.(15) In the current study, in the case of the singledefect of the medial part of posterior table of the frontalsinus adjacent to the frontal ostium, the axillary flaptechnique and agger nasi cells removal were sufficient forexposure and repair.

(Fig. 5) shows how to utilize the new classification inselecting the most appropriate and least destructiveapproach to handle each detailed defective site. In case ofabsence of middle turbinate due to disease or prior

surgery (4 cases, 4.25%), the direct transnasal approachwas adopted.

Although this classification was conducted on onlyprimary cases and few iatrogenic CSF leaks, it could beapplied also in recurrent cases of CSF rhinorrhea. Inrecurrent cases, and in developmental and iatrogenicleaks, which are usually associated with missinganatomical structures e.g. middle and/or superiorturbinate; the approach may need some modification.

Although definite routes are suggested in the currentstudy for each detailed defective site, combination ofapproaches and/or other trajectories may be adopted by

different surgeons according to their experience.

With the application of the suggested new classification inselecting the trajectory followed for defect repair, primarysuccess rate was 93.9%. And secondary success rate was100%.

7/28/2019 2-9 Reda Kamel

7/8



Anterior to BL Posterior to BL

Medial to MTLateral to MT Medial to STLateral to ST

FS AE SS

ACPLat FR FE LL PE LRPSPCP SFMed C

Skull base defects [CSF Leak]

Anterior to BL Posterior to BL

Medial to MTLateral to MT Medial to STLateral to ST

FS AE SS

ACPLat FR FE LL PE LRPSPCP SFMed C


Fig 4. Show s the new classi fi cati on of CSF rhi norrhea accordi ng to detai l ed sit e of skul l base defect. BL = basal l amel l a,

M T = m i ddle turbi nate, ST = superi or t urbinat e, FS = f ront al sinus, AE = ant eri or ethmoid, SS = sphenoid sinus,Lat = l ateral fr ontal sins, M ed = medi al fr ontal sinus, FR = frontal recess, FE = fovea ethmoidal is, LL = l ateral l amel l a,

ACP = anteri or cribri form pl ate, PE = posteri or ethmoid, PCP = posteri or cribri form pl ate, PS = planum sphenoidal e,

SF = sel l ar fl oor, C = cli vus, LR = l ateral recess.

AE

Lat FR FE LL

Transnasal - EndoscopicExternal

Transethmoid

PE SS

LRPS

Transcribriform Trans-PPFOPF

FS CP

PCP

Ttransphenoid

SFMed

Axillary flap

CACPPE


AE

Lat FR FE LL

Transnasal - EndoscopicExternal

Transethmoid

PE SS

LRPS

Transcribriform Trans-PPFOPF

FS CP

PCP

Ttransphenoid

SFMed

Axillary flap

CACPPE


Fig 5. Show s how t o apply t he new classi fi cati on in sel ecti on of the most appropri ate and least dest ructi ve approach t o

t he sit e of t he l eak. FS = fr ontal sinus, AE = ant eri or et hmoi d, PE = posteri or ethmoid, CP = cribri form pl ate, SS =

sphenoid si nus, Lat = l ateral f ront al si ns, M ed = medial front al si nus, FR = frontal recess, FE = fovea ethmoidal i s, LL =

lat eral lamell a, PE = posterior ethmoid, ACP = anteri or cribri form pl ate, PCP = posterior cribri form pl ate, PS = pl anum

sphenoidal e, SF = sel l ar fl oor, C = cli vus, LR = l ateral recess, OPF = osteoplasti c flap, PPF = pt erygopalat i ne fossa. In

case of absence of mi ddl e t urbi nat e due t o di sease or pr i or surgery , the dir ect t ransnasal appr oach w as adopted.

7/28/2019 2-9 Reda Kamel

8/8



CONCLUSIONS

Classification of the skull base defects according to thedetailed site of the leak can be utilized as a roadmap forthe most direct and least traumatic approach for repair. Itpreserves integrity of the sinonasal structures andfunction. It offers precise defect identification, graft

application and support with good results.

Nevertheless, preoperative definition of leaking site iscrucial and its application needs enough sinonasalendoscopic experience.

REFERENCES

1. Ommaya AK. Spinal fluid fistulae. Clin Neurosurg.1976;23:36392.

2. Tabaee A, Kassenoff TL, Kacker A, et al. The efficacy ofcomputer assisted surgery in the endoscopic management ofcerebrospinal fluid rhinorrhea. Otolaryngology - Head andNeck Surg. 2005;133:936-43.

3. Mattox DE, Kennedy DW. Endoscopic management ofcerebrospinal fluid leaks and cephaloceles. Laryngoscope.1990;100:85762.

4. Wormald PJ. Cerebrospinal Fluid Leak Closure. In:endoscopic sinus surgery anatomy, three-dimensionalreconstruction, and surgical technique, Thieme.2005:109-18.

5. Zweig JL, Carrau RL, Celin SE, et al. Endoscopic repair ofcerebrospinal fluid leaks to the sinonasal tract: predictors ofsuccess. Otolaryngol Head Neck Surg. 2000;123:195-201.

6. Wigand WE. Transnasal ethmoidectomy under endoscopiccontrol. Rhinology. 1981;19:7-15.

7. Lopatin AS, Kapitanov DN, and Potapov AA. Endonasalendoscopic repair of spontaneous cerebrospinal fluid leaks.Arch Otolaryngol Head Neck Surg 2003; 129:859-863.

8. Mattox DE, Loury M. Endoscopic management ofcerebrospinal fluid leaks and cephaloceles. In Kennedy DW,Bolger WE and Zinreich SJ editors. Diseases of the sinusesdiagnosis and management. BC Decker Inc. Hamilton,London, pub. 2001:335-40.

9. Hosemann WG, Weber RK, Keerl RE, Lund VJ. Endoscopicsurgery of skull base. In minimally invasive endonasal sinussurgery. Thieme. 2000:76-69.

10. Stankiewicz JA. Cerebrospinal fluid fistula and endoscopicsinus surgery. Laryngoscope. 1991;101:2506.

11. Hegazy HM, Carrau RL, Snyderman CH, et al. Transnasalendoscopic repair of cerebrospinal fluid rhinorrhea: a meta-analysis Laryngoscope. 2000;110:1166-72.

12. Germani RM, Vivero R., Herzallah et al. Endoscopicreconstruction of large anterior skull base defects usingacellular dermal allograft. Am J Rhinol. 2007;21:615-18.

13. Casiano RR, Jassir D. Endoscopic cerebrospinal fluidrhinorrhea repair: Is a lumbar drain necessary? OtolaryngolHead Neck Surg. 1999;121:74550.

14. McMains KC, Gross CW, Kountakis SE. EndoscopicManagement of Cerebrospinal Fluid Rhinorrhea.Laryngoscope. 2004;114:1833-7.

15. Schlosser RJ, Bolger WE. Endoscopic management ofcerebrospinal fluid rhinorrhea. Otolaryngol Clin N Am.2006;39:523-38.

16. Castelnuovo P, Mauri S, Locatelli D, et al. Endoscopic repairof cerebrospinal fluid rhinorrhea learning from our failures.Am J Rhinol. 2001;15:33342.

17. Kirtane MV, Gauitham K, Upadhyaya SR. Endoscopic CSFrhinorrhea closure: our experience in 267 cases. OtolaryngolHead Neck Surg. 2005;132:208-12.

18. Lee TJ, Huang CC, Chauang CC, et al. Transnasal endoscopicrepair of cerebrospinal fluid rhinorrhea and skull base defect:ten-year experience. Laryngoscope.2004;114:1475-81.

19. Lindstrom DR, Toohill RJ, Loerhl TA, et al. Management ofcerebrospinal fluid rhinorrhea: the Medical College ofWisconsin experience. Laryngoscope. 2004;114:969-74.

20. Tabaee A, Anand VJ, Cappabianca P, et al. Endoscopicmanagement of spontaneous meningoencephalocele of thelateral sphenoid sinus, J Neurosurg. 2010;112:10707.

21. Simmen D, Jones N. Management of Skull Base Lesions witha CSF Leak. In Manual of endoscopic sinus surgery and itsextended approaches. Thieme. 2005:240-53.

22. Woodworth BA, Schlosser RJ. Frontal sinus CSF leak. In: DrafW, Senior B, Kountakis S, editors. The frontal sinus.Heidelberg, Germany: Springer. 2005:143-152.

Documents

2-9 Reda Kamel