MIME-Version: 1.0 Content-Location: file:///C:/A3A94A70/CSF-leak-091120.htm Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="windows-1252" CSF Rhinorrhea

TITLE: CSF Rhinorrhea
SOURCE: Grand Rounds Presentation, UTMB, Dept. of Otolaryngology
DATE: November 20, 2009
RESIDENT PHYSICIAN: David Gleinser, MD
FACULTY PHYSICIAN: Patricia Maeso, MD
Discussant: Patricia Maeso, MD
SERIES EDITORS: Francis B. Quinn, Jr., MD
ARCHIVIST:  Melinda Stoner Quinn, M= S(ICS)

=

"This material was prepared by resident physicians in partial fulfillment of educational requirements established for the Postgraduate Training Program of the UTMB Department of Otolaryngology/Head= and Neck Surgery and was not intended for clinical use in its present form. It = was prepared for the purpose of stimulating group discussion in a conference setting. No warranties, either express or implied, are made with respect to= its accuracy, completeness, or timeliness. The material does not necessarily reflect the current or past opinions of members of the UTMB faculty and sho= uld not be used for purposes of diagnosis or treatment without consulting appropriate literature sources and informed professional opinion."

=

Introduction

Cerebrospinal fluid (CSF) rhinorrhea, no matter = the underlying etiology, is the result of an osseous defect at the skull base coupled with a disruption of the dura mater and arachnoid with a resultant = pressure gradient leading to the CSF leak.  = Having knowledge about these leaks, how they occur, and how to treat them is impor= tant for multiple reasons, but none as important as the risk of intracranial infection as they persist.

Physiology

Prior to discussing CSF rhinorrhea, basic knowle= dge of anatomy and physiology of CSF should be reviewed.  Roughly 50-80% of daily CSF is made by = the choroid plexus while the other 30% is produced by the ependmyal surface.  CSF production is the result of capilla= ry ultrafiltration occurring through epithelial cells.  This is regulated through multiple Na+/K+ ATPases.  A Na+/K+<= /sup> ATPase on the vessel lumen side takes Na+ into the cell while another Na+/K+ ATPase on the ventricular side moves t= he Na+ ions into the ventricle.  Water then follows the movement of these ions into the ventricle lea= ding to CSF production.  CSF also consis= ts of multiple other ions (K+, Mg2+, Ca2+, Cl-, and HCO3-) as well as multiple amino acids and proteins, few cel= ls (polymorphonuclear and mononuclear cells), and glucose (roughly 60-80% of blood glucose).   At any one time, roughly 90-150mL of C= SF is in circulation.  The production rat= e has been quoted at being around 20mL/hr and 500mL/day.

Causes

Trauma

There are many causes of CSF rhinorrhea and they= are classified into different categories.  The first category to be discussed is traumatic causes.  This category can be further divided in= to non-surgical and iatrogenic causes.  It should be noted that most traumatic/iatrogenic CSF leaks occur in the area = of anterior cranial fossa since the dura in this area is tightly adherent to t= he thin skull base and easily torn with injury to the skull base.

Non-surgical trauma, whether it is blunt or penetrating trauma, accounts for nearly 80% of all CSF leaks.  Of all the cases of major head trauma, = roughly 2-3% result in a CSF leak.  If a sk= ull base fracture is present a CSF leak is seen in 15-30% of cases.  These leaks may be either immediate (wi= thin 48 hours) or delayed.  Nearly 95% o= f all delayed leaks will manifest within the first 3 months of injury, and are thought to occur as a result of a delayed elevation of intracranial pressur= e (ICP), lysis of clots in and around the region of injury, resolution of edema, or = loss of vascularity with resultant necrosis of soft tissue and bone.

Iatrogenic trauma accounts for 16% of traumatic = cases of CSF rhinorrhea.  In the past, neurosurgical procedures were the most common causes of surgically produced= CSF leaks (resection of pituitary tumors and suprasellar masses).  Although the risk of CSF leak with endo= scopic sinus surgery (ESS) is reported to be around 0.5%, many more surgeons are utilizing endoscopic techniques in practice today.  For this reason, ESS has become the num= ber one cause of iatrogenic induced of CSF rhinorrhea.  The most common site of injury during E= SS is the lateral cribiform lamella, typically on the right side.  Other sites of common injury include the posterior fovea ethmoidalis and the posterior aspect of the frontal recess.=

Atraumatic

The next category to be discussed is non-traumat= ic causes of CSF rhinorrhea.  Non-trau= matic causes of CSF rhinorrhea account for 4% of all cases of CSF rhinorrhea.  This category can be further classified= based on the presence or absence of increased intracranial pressure.

High pressure leaks (increased ICP) account for = 45% of non-traumatic cases of CSF rhinorrhea.  Sustained increased ICP is thought to lead to remodeling and thinning of the skull ba= se that is theorized to be due to ischemia from compression of vessels that ev= entually leads to the formation of a skull base defect.  There are multiple causes of increased ICP with tumor growth and hyd= rocephalus being two of the more common causes. 

Normal pressure leaks represent 55% of non-traum= atic cases of CSF rhinorrhea.  It is wit= hin this category that spontaneous leaks may be placed.  It is hypothesized that the cause of spontaneous leaks are the result of physiologic alterations in CSF pressure= that lead to point erosions in the skull base resulting in a CSF leak.  This theory is based on the fact that i= t has been shown that every few seconds, elevations in ICP up to 80 mmH2O occur.  Other non-traumatic causes include erosion of the skull base by tumors (nasopharyngeal carcinoma, angiofibroma, inverting papilloma, osteomas) and other osteolytic causes (sinusitis, syphilis, and mucoceles).

Another category that will be discussed on its o= wn is the congenital causes of CSF rhinorrhea.  We are discussing them outside of the other categories because congenital causes may be associated with or without increased ICP.  The first, and typically the most commo= n of these defects involve the failure of the closure of the anterior neuropore.  This can lead to the he= rniation of the meninges through the defect (encephaloceles).   Another congenital defect that may res= ult in CSF rhinorrhea is a persistent craniopharyngeal canal.  This is a = vertical midline defect connecting the middle cranial fossa to the sphenoid sinus.  Primary Empty Sella Syndrome res= ults in CSF rhinorrhea in a very similar fashion as other spontaneous CSF leaks,= but contains a congenital component.  E= mpty Sella Syndrome is the condition in which the sella turcica appears empty on imaging.  The primary type is thoug= ht to be secondary to a congenital widening of the diaphragma sella plus another = event that leads to remodeling of the sellar bone with the formation of a skull b= ase defect in that area.  The events th= at have been recognized are an increased ICP being transmitted through the wid= ened diaphragm causing compression of the pituitary.  This is seen with Pseudotumor cerebri, intracranial tumors, and hydr= ocephalus.  The other events include either a ruptu= re or displacement of a cyst into the sella through the widened diaphragm causing compression of the pituitary and resulting in increased pressure within in = the sella turcica.

Evaluation

The initial evaluation of a patient presenting w= ith CSF rhinorrhea begins with a good history and physical.  The typical history is that of clear, w= atery discharge from a single nare.  Ther= e may be an increase in postnasal drip while in the supine position.  The patient may also complain of a salty taste in his or her mouth.  In the = case of headaches, the patient may state that the headache resolves when the leak occurs.  Be highly suspicious when = there is a previous history of meningitis while the leak has been present. 

In most cases, the physical exam is unremarkable= .   However, some maneuvers can be attempt= ed that may point toward a CSF leak.  = First, have the patient lean forward and strain.  This raises ICP and may illicit a leak.  Another way to raise ICP is to compress both jugular veins.   The rhinorrhea is typically clear, but if trauma has occurred, it may be mixed with blood.=   Endoscopic examination of the nasal cav= ities should be part of every physical examination when CSF rhinorrhea is suspect= ed.  Another addition to the physical examin= ation can be determining the presence of the ring or halo sign.  When CSF is mixed with blood, it can be= placed onto a piece of filter paper.  The = blood on the paper will separate out from the CSF (central blood with clear ring = of CSF).  This test is not specific however.   Dula et al found that the ring sign occurred when blood was mixed with water, saline, and other mucus.

Laboratory testing is a very important part of m= aking a diagnosis of a CSF leak.  There a= re a number of tests that can be performed, but only one has been deemed a gold standard in determining whether CSF is present or not, and it is the test f= or beta-2-transferrin.  Beta-2-transferrin is a protein that is produced by enzymes only located in the central nervous system.  The test requires 0.5cc of fluid, and r= esults can be obtained in as little as 3 hours if the test is readably available.<= span style=3D'mso-spacerun:yes'>  However, in most cases the test is requ= ired to be sent out for evaluation, and can take a number of days to return.  All specimens should be refrigerated up= on collecting.  If it is not, the prot= ein will become unstable at room temperature within 4 hours.  If it is refrigerated, the sample is go= od for 3 days.

Glucose testing may be performed, but is not very useful as many things can influence testing.  For example, if blood is present the concentration of glucose will be higher resulting in a false positive.  The presence of meningitis or other intracranial infections will low= er the concentration of glucose in CSF giving a false negative.  If no blood is present, one may suspect= the presence of CSF when the glucose concentration is > 30mg/dL.  If the results show an absence of gluco= se then the specimen does not contain CSF.

Another method for testing glucose has been by t= he use of glucose oxidase paper.  This pap= er changes color with glucose concentrations as low as 5 mg/dL.  This is not ideal as other fluids in the nasal cavity (lacrimal secretions and nasal mucus) contain a similar concentration of glucose, thus false-positive results are common. 

Further testing may include beta-trace protein.<= span style=3D'mso-spacerun:yes'>  This protein is found in CSF, heart, and serum, thus not as specific as beta-2-transferrin in accurately identifying CSF.  The protein levels are also a= ltered in other disease states.   For exam= ple, it is typically elevated in patients with renal insufficiency, multiple sclerosis, cerebral infarctions, and some CNS tumors.  If the serum level of beta-trace protei= n is < 1.0 mg/L, samples with a concentration > 2.0 mg/L may be considered positive for CSF.  However, if the concentration is < 1.5 mg/L, the sample is not likely to contain CSF.  If the test is available, it can be per= formed in 15 minutes.

Imaging

The next step in diagnosing CSF rhinorrhea invol= ves the utilization of imaging.  Most surge= ons would recommend obtaining high resolution computer tomography (CT) scans in= all cases of suspected skull base defects for their ability to detect bony defe= cts.  They can also detect the presence of pneumocephalus, soft tissue masses, and hydrocephalus.  These images should be 1mm in thickness= with axial, sagittal and coronal views.  In add= ition, CT scanning can be combined with the use of intrathecal contrast dye (iophendylate), an imaging modality termed CT cisternography.  This study is more invasive, but is mor= e accurate at identifying the location of a CSF leak. CT cisternography has nearly a 1= 00% detection rate when the CSF leak is active.  However, with intermittent leaks the rate of detection drops to 60%.=

The next imagining modality utilized is magnetic resonance imaging (MRI).  This imag= ing modality is very good at detecting soft tissue abnormalities and distinguis= hing CSF from other fluid located in a sinus cavity as CSF has high signal inten= sity on T2 images.  MRI can also be comb= ined with intrathecal dye injection to improve accuracy.  Overall, this modality is not as good a= s CT when it comes to detecting bony defects and is much more expensive.

Intrathecal tracers

Nuclear medicine testing has also been utilized = in the past to locate the site of a CSF leak, but has fallen out of favor for many reasons.  The study is termed radio= nuclide cisternography.  It involves the in= trathecal injection of a radioactive tracer (technetium-99, I-131, Indium 111).  Prior to this= , pledgets are placed into the nasal cavity in areas suspected of being the site of th= e leak.  The dye is injected, scintigrams of the= skull are obtained, and then the pledgets are removed and measured for radioactiv= e tracer.  The leak site is identified by comparin= g the results of the scintigrams with the pledgets.  However, this study has a lot of drawbacks.  First, it almost always requires an act= ive leak, and even then has been reported at detecting leaks in 70% of cases.  If there is an intermittent leak, the a= bility to pinpoint the area of the leak drops to 30-40%.  Second, positioning of the patie= nt can lead to pledgets in areas away from the leak taking up tracer, thus making = it more difficult to pinpoint the site of the leak.  Third, the radioactive isotope is absorbed into the circulatory syst= em and deposited into normal tissues.

Intrathecal injection of Fluorescein dye is util= ized by many surgeons for identifying the area of a CSF leak preoperatively and intraoperatively.  Typically, an in= jection involving a solution of 0.5%-10% Fluorescein dye is performed.  The patient is then examined roughly 30 minutes to an hour later with an endoscope.  In most cases, the Fluorescein dye can be seen without filters.  However, small defects may only leak a = very small amount only detectable by filters or black light.  In the case of using filters, a yellow = filter is placed over the endoscope while a blue filter is placed over the light source.  It must be noted that it is extremely important to inject low concentrations of Fluorescein as higher concentrations (>500mg of Fluorescein) can lead to severe side effects t= hat include seizures, pulmonary edema, and even death.

Treatment

Medical

The treatment of CSF rhinorrhea can be classifie= d into conservative management and surgical management.  The majority of traumatic CSF leaks res= pond well to conservative management while spontaneous leaks tend to require surgical correction.  In either cas= e, treatment is of extreme importance as the presence of a CSF leak increases = the risk of meningitis 10-fold.

Basic, conservative management revolves around b= ed rest.  The patient is placed on bed= rest for 7-10 days with the head of bed at 15-30 degrees.  The patient is also informed not to str= ain, cough, or perform heavy lifting.  I= t is reported that with this type of management, 75-80% of all traumatic CSF lea= ks will spontaneously resolve. 

The use of antibiotics in the treatment of CSF rhinorrhea remains controversial.  = The reason for their use is to prevent intracranial infection (meningitis).  However, many studies have shown almost= no difference in preventing intracranial infection with or without their use.<= span style=3D'mso-spacerun:yes'>  The fear that many doctors have with th= eir use is the potential to select out more virulent bacterial strains.  Brodie et al published a meta-analysis = in 1997 reviewing 6 studies and 324 patients with CSF leaks.  Two hundred thirty-seven patients were = treated with antibiotics while 87 were not treated with antibiotics.  Meningitis was reported to have occurre= d in 2.5% of patients in the antibiotic group (6/237) and 10% of patients in the no-a= ntibiotic group (9/87). Villalobos et al published a meta-analysis in 1998 that revie= wed 12 studies and 1241 patients with CSF leaks.  Seven hundred nineteen patients were treated with antibiotics while = 522 patients were not treated with antibiotics.  They found that patients were 1.34 times more likely to develop meningitis without the use of antibiotics when a basilar skull fracture had resulted in a CSF leak.  With all c= auses of CSF leak, patients were only 1.10 times more likely to develop meningitis without the use of antibiotics.  Fo= r this reason, they recommended not using antibiotics when CSF leaks are present.<= /p>

When there is increased ICP, diuretic use should= be considered.  Acetazolamide is a diu= retic that inhibits the conversion of water and CO2 to bicarbonate and= H+.  The loss of H+ slows the act= ion of Na+/K+ ATPases that are responsible for the productio= n of CSF, thus lowering ICP.

Another option to consider is the use of a lumbar drain.  In conservative cases, its = use should be considered when a leak fails to responds after 5-7 days of conservative management.  Continuous drainage is recommended over intermittent drainage as this helps to prevent spikes in C= SF pressure.  The rate of drainage sho= uld be about 10-15cc/hr to prevent side effects such as headaches, nausea, and emesis.  There have been cases of patients slipping into comas with too much drainage.

Surgical Treatment

The surgical management of CSF rhinorrhea is categorized into intracranial and extracranial approaches.  Intracranial approaches are utilized wh= en there is a comminuted skull fracture with displacement of fragments requiri= ng reduction, extensive skull base fractures, and fractures associated with intracranial hemorrhages or contusions that ordinarily would require cranio= tomy for treatment.  The defects can be repaired by primary closure with or without the use of grafts.  Some examples of various grafts utilized include free or pedicled periosteal or dural flaps, muscle plugs, mobilized portions of the falx cerebri, fascia grafts, and many commercial grafts.  In addition to this, many surgeons will= reinforce these grafts with fibrin glue to provide a better seal.  The advantages of an intracranial appro= ach is that it provides direct visualization of the defect, allows for the inspect= ion of the adjacent cerebral cortex for injuries, and allows a better chance of patching a defect, especially in the face of increased ICP.  The disadvantages, however, are increas= ed morbidity, increased hospital time, risk of injury to the brain from retrac= tion (hematoma, seizures, cognitive dysfunction, risk of permanent anosmia), and poor visualization of the sphenoid sinus.

The extracranial approach to repairing CSF leaks includes open and closed approaches.  For the purpose of this discussion we will focus on the closed approach utilizi= ng endoscopic techniques, as the majority of surgeons utilize this approach as their first line of surgical treatment since studies have shown that the success of repair is > 90%.  In addition, an endoscopic approach allows for better magnified and angled visualization, does not require external incisions, and minimizes intranasal mucosal injury. 

The key to endoscopic repair of a CSF leak is go= od visualization and exposure of the defect.  Initially, the surgeon must assess whether an encephalocele is prese= nt.  If an encephalocele is present, it shou= ld be cauterized at its stalk with bipolar cautery prior to reduction into the anterior cranial fossa to prevent intracranial hemorrhage.  For good exposure, the surgeon should e= levate the surrounding mucosa to provide 2-5mm of bone exposure around the defect.=   Any mucosa remaining in the defect shou= ld be removed prior to repair to help prevent future mucocele formation.

There are many types of grafts utilized, but it = should be noted that the graft should be roughly 30% larger than the defect to acc= ount for post-operative shrinkage.  The = types of grafting material utilized are cartilage, bone (septum, mastoid tip, mid= dle turbinate), mucoperichondrium, septal mucosa, turbinate mucosa and/or bone,= fascia (temporalis, fascia lata), abdominal fat, and pedicled septal or turbinate flaps.  It should be noted that ped= icled flaps tend to tint, fold, and contract when utilized.

Grafting techniques can be categorized into over= lay, underlay, and combined techniques.  In an overlay technique, a graft is placed directly over the defect.  In an underlay technique, the graft is = placed between the dura and bony defect.  = The combined technique utilizes both underlay and overlay grafts.  In addition to these techniques, most surgeons reinforce the repair with fibrin glue to provide an improved seal.=   The placement of absorbable (gelfoam) a= nd/or non-absorbable packing can further improve the seal.  Most surgeons have moved away from non-absorbable packing since it must be removed post-operatively and may le= ad to the displacement of the graft when removed.  In the face of increased ICP, it is recommended that a multilayered graft be utilized.

The size of the defect also plays a role in the grafting technique utilized.  If the defect is < 2mm, the type of grafting technique utilized typically does = not make much difference as most techniques will be successful in repairing the= CSF leak.  If the defect is 2-5mm, one = must note whether comminuted bone segments or significant dural injury is present.  If they are not present, = the use of an overlay grafting technique is sufficient.  However, if either is present, one shou= ld utilize a composite graft or a separately harvested bone plus mucosa grafti= ng technique where the bone is placed in an underlay fashion while the mucosa = is placed in an overlay fashion.  If the defe= ct is >5mm, the repair should be performed with a composite graft or separate bone plus= mucosa grafting technique as described above. 

Post-operatively, the patient should be placed o= n bed rest with the head of the bed set at 15-30 degrees for 3-5 days.  The patient’s blood pressure should be maintained at a normal level.  The patient should also be placed on stool softeners to prevent straining, and = be instructed to not cough, blow his or her nose, and avoid any heavy lifting.  Some surgeons will utiliz= e a lumbar drain post-operatively.  It = is lift in place for 3-5 days with a maximum drainage of 10-15cc/hr.  If non-absorbable packing is utilized, antibiotics should be given.

Conclusion

In conclusion, CSF rhinorrhea has many etiologie= s, the most common being trauma.  There ar= e many ways to detect the presence of CSF and pinpoint the site of the skull base defect.  Treatment may be either conservative or surgical.  Remember= , CSF rhinorrhea must be recognized and treated appropriately as the risk of intracranial infection is increased 10-fold as it persists.

Discussion by Patricia Maeso, MD of Dr. Gleinser’s G= rand Rounds presentation on CSF leak 11/20/2009=   

Did you go into the primary reason why we decide to close these defects?  What do you think is the reason we take= these patients to the O.R. and try to close these defects?<= /p>

(Dr. Gleinser): “To prevent infection.”

And do you know the risk of meningitis per year?  Some studies quote up to 19% per year.  So that’s really the reason why you take these patients to the O.R.<= span style=3D'mso-spacerun:yes'>  Of course the rhinorrhea is bothersome = but really it’s the meningitis.

I’m a big proponent of vaccinating these patients with the pneumococcal vaccine.  As soon as I know that th= ey have a CSF leak and I’m prepping them for surgery I’ll give them a vaccine right there while they’re seeing me.

You mentioned bone, you mentioned cartilage, you mentioned mucosa.  Fat is also very good to plug up holes.=   And then over it you can put a thin muc= osal graft.

I really like the free graft as well unless it’s really something big.  The pedicled flaps are a workhorse; the= y can be misplaced more frequently than your free mucosal graft due to contractio= n.

I agree with the author you cited that folding and contracture is an issue wi= th that.  For fat grafts I just go to = the belly.  The earlobe is just a littl= e bit tedious.  You must know preoperativ= ely how small your defect is.  Going to= the belly is so much easier plus the quality of the fat is very different.  It just doesn’t disintegrate as easily = and you can get the adipose tissue complete.

I always make sure with a free graft that I always mark it with a marking pen and I paint one side of the free graft so that there’s no confusion as to the muc= osal side.  You can take your free graft= from the septum (a full free graft from the septum), you can take it from the fl= oor and that one is really nice.  The o= ne that I’m tending to like right now is a free graft off of the inferior turbinate because there’s so much that you can get off of it.

I’ve moved away from using non-resorbable packing though it’s really hard and st= able and keeps things in place but it might pull things with it.  I like re-packing it nicely with Gelfoa= m.

The patient should be at bedreset for three to five days and should not sneeze = or blow their nose.  If the patient sh= ould decide to sneeze loudly or blow their nose they’ll blow their graft out.

Whether we use a lumbar drain depends on how big the site of the defect is, how high the ICP is, the leak flow rate, etc…just make sure you think about these things.

Sources:

1.      Welch, KC, MD, and Stankiewicz, J, MD.= CSF Rhinorrhea. eMedicine from WebMD. Online[Available]: http://emedicine.medscape.com/article/= 861126-overview, 2009.

2.      Greenburg, J, MD. Cerebrospinal Fluid Rhinorrhea. Baylor College of Medicine: Department of Otolaryngology. Online[Available]: http://www.bcm.edu/oto/grand/120398.html, 1998.

3.      Ommaya AK. Spinal Fluid Fistulae. Clinical Neurosurgery, 1976;23:363= -392

4.      Cummings, CW, MD et al., eds. Cummings Otolarnygology: Head and Neck Surgery. 4th = ed. 4 vols. Philadelphia: Elsevier-Mosby, 2004.

5.      Briscoe, M, MD. Endoscopic Repair of C= SF Rhinorrhea. UTMB: Department of Otolaryngology. Online[Available]: http://www.utmb.edu/otoref/Grnds/CSF-rhinorrhea-061115/CSF-rhinorrhea-06111= 5.htm, 2006.

6.      Shields, G, MD. Congenital Midline Nas= al Masses. UTMB: Department of Otolaryngology. Online[Available]: http://www.utmb.edu/otoref/Grnds/Nasal-mass-021106/Nasal-mass-021106.htm, 2= 002.

7.      Kizilki= lic, O, MD et al. Hypothalamic Hamartoma Associated with a Craniopharyngeal Canal. American Journal of Neuroradiology, 2005;26:65-67.

8.      Tsai, E= , MD et al. Tumors of the Skull B= ase in Children: Review of Tumor Types and Management Strategies. Neurosurgical Focus, 2002;12:5.

9.      Elias, = MA, MD. Empty Sella Syndrome. CNS Clinic-Jordan. Online[Available]: http://pituitaryadenomas.com/emptysella.htm, 2005= .

10.  Dula, DJ, MD and Fales, F, MD. The 'Ri= ng Sign': Is It a Reliable Indicator for Cerebral Spinal Fluid? Annals of Emergency Medicine, 1993;22:718-720.

11.  Moyer, P. Beta-Trace Protein Shows Pro= mise as a Marker for Diagnosing CSF Leaks. Doctor’s Guide. Online[Available]: http://www.docguide.com/dg.nsf/PrintPrint/5DF097A1EB04B3FA85256C3E00731E65, 2002.

12.  Lemole,= GM, MD et al. The Management of Cranial and Spinal CSF Leaks. Barrow Quarterly. Online[Available]: <= span style=3D'mso-bidi-font-family:Arial'>http://www.thebarrow.org/Education/Bar= row_Quarterly/Vol_17_No_4_2001/162074, 2001.

13.  Villalobos T, MD et al. Ant= ibiotic prophylaxis after basilar skull fractures: A meta-analysis.  Clinical Infectious Diseases, 27:364-369, 1998.

14.  Brodie HA. Prophylactic Antibiotics for Posttraumatic Cerebrospinal Fluid Fistulae= . A meta-analysis.  Archives of Otolaryngology Head Neck Surgery,  123:749-752, 1997.