CSF Rhinorrhea: 95 Consecutive Surgical Cases with Long Term Follow-Up at the Mayo Clinic
Holger G. Gassner, M.D., Jens U. Ponikau, M.D., David A. Sherris, M.D., and Eugene B. Kern, M.D.
A persistent cerebrospinal fluid (CSF) leak is potentially lethal, and surgical treatment is often required. CSF leak repair is an infrequently performed procedure, and only limited information is available on the long term success of the surgical techniques that are used. This retrospective chart review includes 95 patients who underwent various types of repair surgery fin- CSF rhinorrhea at the Mayo Clinic. The purpose of this study was to extract factors such as the choice of sealing material, etiology, location of defect(s), surgical approach, and previous procedures, and to analyze their association with the long term success and failure of surgical repair. The mean time interval iii this study between unsuccessful surgery and recurrence was 50.8 months, and the wean follow-up 109 months. Among the various approaches, defects repaired endonasally had the lowest recurrence rate. Local nasal mucosa advancement flaps failed more frequently (83.3% failure) than other- types of graft material (p = 0.023). These failures took place in a delayed fashion (mean interval until failure: 80 months). Local osteo-mucoperiosteal or chon-dro-mucoperichondrial flap flaps (22.2% recurrence rate) and free graft material (15.6% recurrence rate) had the best outcome. The use of fibrin glue to fixate free grafts did not improve the result in this series. Transcranial procedures were associated with a higher complication rate than extracranial procedures
From the Department of Otorhinolaryngology, Mayo Clinic, Rochester , Minnesota
Presented at the Spring Meeting of the American Rhinologic Society. Palm Desert , California , April 25-26, 1999 Address correspondence and reprint requests to Dr. David A. Sherris, Department of Otorhinolaryngology, Mayo Clinic, 200 First Street SW , Rochester , MN 55905
(12.9% versus 3.2%). Overall, successful repair was achieved in 91.6% of the patients. We discourage the use of mucosa advancement flaps and advocate free grafts or pedicled osteo-mucoperiosteal or chondro-mucoperichonalriul flaps as sealing material of choice in the majority of cases. The occurrence of delayed failure has to he considered when evaluating reports of CSF rhinorrhea after surgical repair. (American Journal of Rhinology 13, 439-447, 1999)
In the second century AD, 6 centuries after the death of Hippocrates, Galen was the primary force in the ancient school of Greek medicine. His science, known as Galenism, remained unchallenged for more than 1000 years. Therefore, not surprisingly, his notion that the release of cerebrospinal fluid (CSF) through the nose was a physiologic process carried into the Renaissance and was not dispelled until the seventeenth century.
The first comprehensive description of CSF leak in a child with hydrocephalus was given by a surgeon, Charles Miller, in 1826.' In 1899, Sir St Claire Thomson of the Royal Ear Hospital , London , reviewed a series of CSF leaks of various etiologies and coined the term "cerebrospinal rhinorrhea. This marked the beginning of nearly a century of rapidly evolving therapeutic concepts resulting in a dramatically improved prognosis for this life-threatening condition.
The history of surgical therapy for CSF rhinorrhea began with the first successful intracranial repair by the American surgeon Walter Dandy in 1929.3 It took another 19 years until the first extracranial repair was performed by the Swedish physician, Gösta Dohlman.4 He repaired a leak of the roof of the ethmoid via a nasoorbital incision. Oskar Hirsch of Austria , one of the early pioneers of trans-sphenoidal pituitary surgery since 1910, was the first to repair a CSF leak in the sphenoid sinus using the transseptal approach in 1952. Donald Vrabec and Erik Hallberg per-formed the first endonasal approach at the Mayo Clinic in 1964.7. The latest milestone in CSF repair surgery was added by Malte Wigand in 1981. who introduced the intranasal endoscopic technique at the University of Erlangen Germany .
CSF repair surgery is not frequently reported. and relatively few studies have been published examining a large number of consecutive series. A variety of surgical techniques were used, and reported results were very favorable. Yet most series differ significantly in important prognostic factors like the duration of the postoperative follow-up. Results are therefore difficult to compare. and the prognostic importance of factors like the choice of sealing material for the various repair techniques have not been clearly established.
The objective of this article is to determine the characteristics or factors related to successful CSF leak repair in a large group of consecutive patients treated by various surgical techniques.
MATERIALS AND METHODS
Patient Selection Cases of CSF rhinorrhea of otologic origin were not analyzed. A total of 120 consecutive cases of CSF rhinorrhea of rhinologic origin were treated surgically at the Mayo Clinic from 1975 through 1997. Among these 120 patients, the following inclusion criteria were applied:
Only patients who approved review of their charts were included. Successful cases were only included if postoperative follow-up was longer than 12 months. All surgical failures were included, even if postoperative follow-up was less than 12 months. Ninety-five of the 120 patients met the inclusion criteria and became subjects of this retrospective chart review. Eighty-two of the 95 patients had primary repairs at the Mayo Clinic, whereas for 13 patients the first procedure at the Mayo Clinic was a revision repair.
Data collected included etiology, location, type of repair. graft material used, need for revision, and complications. Additional follow- up information was established with a questionnaire sent to the patients. If questionnaires were not answered, patients were interviewed by telephone.
The etiology of the CSF rhinorrhea was classified as follows: 1. Traumatic origin: History of a significant trauma to the head and/or obvious findings at surgery consistent with a traumatic origin; iatrogenic causes excluded.
Neurosurgical origin: History of a previous intracranial intradural procedure (such as removal of a pituitary adenoma). Iatrogenic leaks, which were closed during the same intracranial procedure. were not included.
Rhinosurgical origin: A paranasal sinus or other external or endonasal extradural procedure (such as orbital decompression).
Spontaneous: Obvious causes of nontraumatic CSF rhinorrhea, such as congenital malformations or tumors and/or nothing in the history or at surgery consistent with a traumatic or iatrogenic origin.
The site of the identified intraoperative leak was classified into the following categories: I. Sphenoid sinus: one or more leaks confined to the sphenoid sinus.
Ethmoid sinus: one or more leaks confined to the anterior and/or posterior ethmoid air cell system.
Cribriform plate: one or more leaks confined to the cribriform plate.
Frontal sinus: one or more leaks confined to the frontal sinus.
Multiple: more than one leak in more than one of the above mentioned sites or one or more leaks extending over more than one of these sites.
The surgical approaches to the leak site were classified as follows:
Transfacial extracranial approach
Transcranial intradural or extradural approach
Any combination of the above
The graft material used for repair of the defect was classified as follows: I Free graft: one or a combination of the following free graft materials: autologous muscle, fat, nasal mucosa. fascia. or lyophilized dura without the addition of fibrin glue or fixation by sutures.
Free graft with fibrin glue added: free graft as de-scribed in (I) fixed with fibrin glue.
Composite grafts. These include a combination of free graft material as described in ( 1) with autologous cartilage or hone. and free chondro-mucoperichondrial or osteo-mucoperiosteal composite grafts.
Local flaps, including nasal mucosa advancement flaps.
Reinforced nasal mucosa flaps. This category comprises local flaps as described in (4) covering underlying autologous free graft material such as muscle or cartilage and pedicled osteo-mucoperiosteal or chondro-mucoperichondrial flaps.
Other: graft material not classified in any of these groups, such as suture fixated dura mater grafts.
The success rate of the first and subsequently numbered repair surgeries was analyzed for operations per-formed at the Mayo Clinic only. Success and failure of previous surgical procedures performed elsewhere was not counted. Failure of an operation was defined by the occurrence of one of the following criteria at any time postoperatively:
I. Recurrence of CSF rhinorrhea
Radiographic diagnosis of intracranial air
Clinical and laboratory diagnosis of meningitis Overall success was considered the percentage of patients who were eventually treated successfully, even if multiple surgeries were necessary.
Associations between failure (recurrence) after first surgical repair and etiology, location, and surgical approach were evaluated based on Fisher's exact test. All calculated p-values were two-sided and p-values less than 0.05 were considered statistically significant. The failure rate after first surgical repair was used as a criterion to quantify the prognostic value of these factors. Failures within the first 12 months after repair were categorized as early failures, while those occurring more than 12 months after repair were called late failures. The size of the defect was not consistently recorded in the records. Therefore, this potential confounding factor was not controlled in this study.
Clinical and/or radiographic and/or laboratory diagnosis of any one of six major complications that occurred any time during the postoperative period could include meningitis, tension pneumocephalus. intracerebral abscess, intracerebral hemorrhage, blindness, and death.
The mean age of our 95 patients was 41.1 years, with a range of 4 years through 86 years. Fifty-one percent of the patients were female (48/95). 49% (47/95) were male. Sixteen patients have died from unrelated causes.
The average follow-up was 109 months, with a range of I month to 273 months. Repairs were classified as successful only if follow-up was at least 12 months.
The most frequent etiology in this series of 95 consecutive patients was a neurosurgical procedure with 49.5% of the cases (47/95), followed by trauma with 242% (23/95). spontaneous 14.7% (14/95), and a rhinosurgical procedure with 11.6% (11/95) (Fig. 1).
The most frequent site of the skull base defect was the sphenoid sinus with 49.5%% of the cases (47/95), followed by the cribriform plate with 17.9% (17/95), the ethmoid air cells with 16.8% (16/95), and the frontal sinus with 4.2% (4/95) of the cases. Multiple sites were seen in 11.6% of the patients (11/95) (Fig. 2).
First Surgical Approach
Of the 95 patients. 13 patients were referred from else-where for revision surgery of their recurrent CSF rhinorrhea. The surgical approach used for primary repair surgery at the Mayo Clinic was analyzed for the remaining 82 patients. Thirty-eight percent (31/82) of these patients underwent a transseptal approach. 31.7% (26/82) a transfacial approach. and 20.7% (17/82) a transcranial approach. An endonasal approach was performed in 8.6% (7/82). and a combined approach in 1.2% (1/82) of these cases (Fig. 3).
A spinal tap was performed to decrease cerebrospinal fluid pressure preoperatively in five first surgical cases. Two of these cases were successfully closed and the CSF leak was stopped; three recurred.
In order to identify the site of the leak, an additional eight lumbar taps were performed for the injection of fluoresce-in. A total of 0.5-0.75 mL of 5% fluorescein was diluted into 5 mL cerebrospinal fluid and then injected into the lumbar subarachnoid space. In our series, there were no adverse reactions from the introduction of 5% fluorescein. The methods, limits, and possible complications of this diagnostic test in 925 cases were reported by Wolf et al."
Eighty-two of the 95 patients underwent their primary repair surgery at the Mayo Clinic. Seventy-three percent (60/82) of these first repairs were performed successfully.
The success rate dropped to 58.6% (17/29) for second repairs and 57.1% (8/14) for third repairs. Of the fourth operations, 66.7% were successful (2/3 fourth repairs). A total of 129 operations were performed. with a success rate of 67.4% (87/129) per operation.
Of the 95 patients in this series. 87 patients (91.6%) were eventually treated successfully. Among the 8 patients who could not he cured of their CSF rhinorrhea, two declined further surgery after one and three attempts. respectively. One patient with Dandy-Walker syndrome also could not he cured. The patient's malformations included left anophthalmos with severe hypoplasia of the left face and skull, as well as a communicating hydrocephalus. Another long-term failure occurred after a subtotal removal of a frontoparietal parasagittal meningeoma with subsequent sagittal sinus thrombosis. The remaining failures followed orbital decompression surgery. trauma with four previous attempts at closure elsewhere, an empty sella syndrome, and a congenital malformation of the skull base.
These data include one patient who presented with men
Neurosurgery (47/95) 49.5%
Rhinosurgery (11/95) 11.6%
Spontaneous (14/95) 14.7%
Trauma (23/95) 24.2%
Sphenoid Sinus (47/95) 49.5%
After revision CSF repair surgery and placement of a right ventriculoperitoneal shunt for increased CSF pressure. No evidence for recurrent CSF rhinorrhea was found in this patient. Therefore, the CSF repair surgery in this patient was classified as successful, and the cause of his meningitis was attributed to his ventriculoperitoneal shunt. All other repairs that were followed by meningitis, pneumocephalus, or recurrent CSF rhinorrhea were classified as unsuccessful with recurrence (Tables I and I I ).
Increased CSF Pressure
Eight patients in this series were diagnosed with in-creased CSF pressure preoperatively or intraoperatively. Seven of these patients with high pressure leaks were treated with the placement of a ventriculoperitoneal shunt before or during their CSF repair. Three of these patients were treated successfully at first repair and three at the second repair surgery. The defects of the remaining two patients with high pressure leaks could not be closed successfully.
Outcome of 129 Operations for CSF Rhinorrhea
Performed at the Mayo Clinic in 95 Patients
Number of Operations
Success Rate (%)
First repair surgery
Second repair surgery
Third repair surgery
Fourth repair surgery
Fifth repair surgery
Overall Outcome in 95 Patients Operated for CSF
Number of patients
Failure Related to First Surgical Approach
The recurrence rate after the first surgical procedure performed at Mayo (n = 82) was used as a criterion in order to determine the relative failure rate of the various surgical approaches. The lowest recurrence rate of 14.3% (1/7) occurred in the endonasal approach. The intracranial approach yielded the second best result with 17.6% (3/17) recurrence, followed by the transseptal approach with 25.8% (8/31) recurrence, and the transfacial approach with 34.6% (9/26) recurrence. Once, a leak was repaired through an intracranial and transfacial approach simultaneously. This combined approach failed. The association between recurrence and type of surgical procedure was not statistically significant (Fig. 4).
Failure in Relation to the Choice of Graft Material
Failure in 82 first repair surgeries at Mayo Clinic was compared by the choice of graft material. Based on the Fisher's exact test. a statistically significant association between type of graft material and failure rate (p = 0.023) was found. Free graft material such as autologous muscle, abdominal fat, or free nasal mucosa was used for 32 of the 82 first repairs and achieved the best outcome, with a failure rate of only 15.6/ (5/32). Reinforced local mucosa flaps or pedicledosteo-osteo-mucoperiosteal or chondro-mucoperichondrial flaps were used nine times. This sealing material had the second lowest failure rate with 22.2% (2/9) recurrence. Composite grafts such as free osteo-mucoperiosteal grafts. free chondro-mucoperichondrial grafts. or other free graft material reinforced by underlying cartilage or bone failed in 25.0% (5/20) of cases. Fibrin glue was used 11 times to fixate free autologous graft material. Such closure failed in 27.3% (3/I I) of the cases. Advancement flaps of nasal mucosa only were used six times. All but one (83.3%) of such closures ultimately failed. This failure rate is statistically significantly higher than that of the other categories. In four cases the sealing material was classified as other graft material and not analyzed for failure rate (Fig. 5).
Surgical Failure in Relation to the Location of the Defect
Failure at the first surgery was most common for defects in multiple sites (50.0% failure, 3/6 cases). followed by defects of the sphenoid sinus (28.6%; 12/42 cases), the cribriform plate (23.5%, 4/17 cases), and the ethmoid sinus (23.1%. 3/13 cases). CSF rhinorrhea originating in the frontal sinus was repaired without failure in 100% (4/4) of the cases. The association between surgical failure and location of the defect did not attain statistical significance (Fig. 6).
Surgical Failure Related to Etiology
First repair of spontaneous CSF leaks failed in 38.5% (5/13) in this series. Traumatic leaks recurred in 31.3% (5/16) of the cases. Dural leaks subsequent to a rhinosurgical or neurosurgical procedure had the lowest failure rate with 22.7% (2/9) and 22.2% (10/44), respectively. This association was not statistically significant (Fig. 7).
Eight of 95 patients (8.4%) had one or more complications postoperatively. No intracranial hemorrhage, blindness. or death was caused by CSF leak repair surgery. Five patients (5/95) developed a tension pneumocephalus. Two of these patients (2/5) had meningitis, and one (1/5) had an intracerebral abscess at the same time as the pneumocephalus. Meningitis alone was diagnosed in three patients (3/95). An intracranial approach had been used in four (4/8). an extracranial approach in three (3/8), and a combined approach in one (1/8) of these eight patients with complications.
Six of the eight patients with complications required a revision surgery or an additional procedure, such as abscess drainage or decompression of a tension pneumocephalus.
Time Span until First Recurrence
In 21 of these 29 patients who underwent second repair surgery at Mayo, the mean interval between primary surgery and first recurrence could be determined. This interval was 50.8 months (range: l day-28 years). Nine of these failures occurred within the first postoperative year, whereas the remaining 12 failures occurred later than I year after the first surgery.
The present and previous series differ noticeably in important factors such as age and gender distribution, etiology, surgical technique, and especially postoperative follow up. The size of the defects as a potential confounding factor could not be analyzed in the present study. Although the relatively large sample size makes a skew of the data unlikely, this limitation has to be considered when evaluating the present results. The average follow-up of this study was 109 months (range 1 month to 273 months) with a minimal follow-up for successful cases of 12 months. The majority of failures were delayed (later than I year after repair), resulting in a mean time interval until the first recurrence of 50.8 months. The importance of pro-longed postoperative follow-up cannot be overemphasized when evaluating results of CSF leak therapy. In fact, only a limited number of large series was found in the literature, with a postoperative follow-up that exceeded the mean time until first failure in this series. Eventually 91.6% of the patients in this series were treated successfully. Among the eight patients who were not treated successfully. two had declined revision surgery. The remaining six complex atypical cases would he expected to have a higher rate, of failure.
Complications occurred in eight patients for a rate of 8.4% (8/95). All but one of these complications resulted from revision repairs, and an intracranial or combined approach had been used in five of these eight cases. The complication rate of all procedures that required a transcranial approach was higher (12.9%) than that of all extracranial operations (3.2%) (p = 0.063). The lower complication rate of extracranial versus intracranial approaches makes the extracranial approach appealing to consider for first time leak cases.
The success rate per operation decreased with the number of revision repairs. First repair was successful in 73.2% of the cases, second and third in 58.6% and 57.1 %, respectively. These data, as well as the possible detrimental sequelae of a persistent dural leak, have to he considered when counseling a patient who presents with recurrence of CSF rhinorrhea after surgical treatment.
Defects in multiple sites had the highest failure rate after first repair surgery with 50% (3/6) recurrences. A considerable number of leaks in this series was located in the sphenoid sinus. The majority of these patients had leaks secondary to transseptal trans-sphenoidal hypophysectomy. Closures of sphenoid leaks failed in 28.6% (12/42) of the cases, followed by leaks of the cribriform plate with 23.5% (4/17) recurrence, and the ethmoid sinus with 23.1% (3/13) recurrence. All leaks (4/4) located in the frontal sinus were closed successfully. These data suggest that leaks located in more unfavorable sites such as the sphenoid sinus or the cribriform plate are more prone to failure than leaks in a rather favorable location, like the frontal sinus.
Based on success at first repair, spontaneous CSF leaks had the poorest outcome with a failure rate of 38.5% (5/13). Spontaneous CSF leaks, according to the scheme developed by Ommaya, are those of nontraumatic etiologies. This category includes causes like hydrocephalus, congenital skull base defects, and tumors. We think it is reasonable to assume that a part of the failures in this group may be attributed to persistence or recurrence of the underlying disorder. CSF leaks of traumatic origin had the second poorest repair outcome, with 31.3% (5/16) failure. Multiple other injuries, increased frequency of an intracranial procedure, cerebral edema with increased intracranial pressure, or the possibly larger size of the defect may be contributing factors. On the other hand iatrogenic leaks could be repaired with a failure rate of 22.6%. The nature of the initial procedure that caused the leak, rhinosurgical or neurosurgical, did not affect the outcome (22.2% vs. 22.7%).
Success rates for intracranial and extracranial CSF leak repairs have been reported to be in similar ranges. In this series, the average intracranial repair was more successful with 17.6% recurrences, compared to the average extracranial repair with 28.1% recurrences. Although the site and size of the defect frequently dictate the choice of the surgical approach to the dural leak, the availability of endoscopes and, more recently, computerized image-guided systems have made an increasing number of CSF repairs accessible by the endonasal surgical approach. The endonasal approach was the single most successful approach, with 14.3% recurrence rate. However, only seven endonasal approaches could be included in this series.
The first concern of the surgeon deciding on the right graft material should be the long-term success of the operation. However, other factors such as preserving nasal physiologic function by avoiding injury of the nasal mucosa. donor site morbidity. and underlying disease have to be taken in account as well. In this series, all but one (5/6) of the local mucosa advancement flaps failed. In theory, this viable tissue should heal well because its perfusion is maintained. However, after a mean time interval of 80 months. 83.3% (5/6) of these flaps had failed. Hosemann elegantly demonstrated shrinkage of free mucosa transplants of about 1/5 in the early postoperative phase in rabbits.24 Our hypothesis is that over time, contraction of the pedicle continues to increase tensile forces, which eventually causes a disruption of the initially healed defect.
On the other hand, reinforced local flaps had a low recurrence rate of 22.2% in 11 cases. Yessenow reported success in 14 cases of osteo-mucoperiosteal flaps with a follow-up of more than 7 years. These data suggest that additional underlying material either plugs the defect itself or helps to withstand the contractile forces of the pedicle. In light of these findings, we advocate considering alternatives to the use of local mucosa flaps. such as free autologous grafts or reinforced flaps.
The use of fibrin glue for CSF leak repair was first reported in 1981.8 In our series, free grafts fixated with fibrin glue failed more frequently than those without fibrin glue. Other authors report routine use of fibrin glue in CSF repair surgery for the fixation of free transplants. In a rat model, Nishihira and McCaffrey showed that the success rate of free autologous muscle in combination with fibrin glue was better than that of free autologous muscle alone.
However, this finding was not statistically significant. As the number of uses of fibrin glue with free graft material in the present series is not large, and the difference of the failure rates between repair with and without fibrin glue (27.3% versus 15.6%) did not attain statistical significance, we hesitate to discourage use of fibrin glue with free graft material. However in light of the good results in a large number of cases, we recommend the use of free grafts alone as an excellent option for the repair of applicable defects.
The outcome of free composite grafts with 25.0% (5/20) failure has to be seen in light of the size of the defect. It is reasonable to assume that larger defects are more likely to recur. Unfortunately, the size of the defect was not always recorded in the charts: and the influence of defect size on the prognosis of CSF leak surgery has yet to be determined in future studies. We feel that the present data encourage the use of cartilage or bone in order to stabilize larger defects.
The two most successful sealing materials for repair of dural leaks in this series were free grafts. followed by reinforced nasal mucosal flaps. Such flaps may cause disturbed nasal physiology or extensive injury to the nasal mucosa, with a long-term risk of atrophic rhinitis and its associated morbidity. Since Moore et al. have demonstrated that extensive injury to the nasal mucosa can have a detrimental effect on the physiology of the nasal cavity, we urge caution when mobilizing nasal mucosa.'' We therefore advocate preservation of the nasal mucosa whenever possible. Hence, we favor free grafts. e.g., autogenous muscle, temporalis fascia, or fascia lata as an excellent sealing material for CSF repair surgery.
The endonasal endoscopic technique and the use of image-guided systems is preferred in our department. The use of nasal mucosa advancement flaps alone has been discarded. It is our experience that the B-2-transferrin test is an excellent screening tool in the diagnosis of CSF rhinorrhea, and visualization with intrathecal flurescein is an excellent method of localizing the site of the leak.
Ninety-two percent of the patients in this series were eventually treated successfully. The majority of the unsuccessful cases presented with severe underlying conditions or declined further repair. We conclude that it is reasonable in the majority of cases to proceed with revision surgery rather than risk development of meningitis in the patient with a recurrent or a persistent CSF leak.
The graft material of choice should he a free autogenous graft or a pedicled osteo-mucoperiosteal or chondro-mucoperichondrial flap. The use of fibrin glue to fixate free graft material did not improve the outcome.
We emphasize the importance of determining the nature of the underlying pathologic condition causing a spontaneous CSF rhinorrhea.
The site of a defect seems to be an important prognostic factor of CSF repair surgery. Repair of defects in more favorable sites showed a trend toward a better outcome than repair of defects in less favorable sites.
Future studies are necessary to analyze the prognostic implication of additional factors, like the size of the skull base defect.
The use of mucosa advancement flaps alone in CSF repair surgery is discouraged because of probable contracture over time and eventual recurrence of CSF rhinorrhea.
Considering the long average interval until failure (50.8 months) with recurrence, reports of success with surgical repair of CSF leaks have to be evaluated in light of a sufficient postoperative follow-up, at least 1 year. Four years or longer is preferred.
The complication rate was 8.4% (8/95 cases), including tension pneumocephalus, meningitis, and intracerebral abscess. There were no cases of death, blindness, intracranial hemorrhage, or abscess.
The authors acknowledge the Department of Neurosurgery for their magnificent contribution and compassionate care of our mutual patients.
- Miller C. A case of hydrocephalus chronicus, with some unusual symptoms and appearance on dissection. Trans Med Chir Soc Edinburgh 2:243-248. 1826.
- Thomson St C. The Cerebro-Spinal Fluid. Its Spontaneous Es-cape from the Nose. London, Cassell and Co., Ltd.. 8, 1899.
- Dandy W. Pneumocepahlus (intracranial pneumatocele or aerocele). Arch Surg 12:949-982. 1926.
- Dohlman G. Spontaneous cerebrospinal rhinorrhea. Acta Otolaryngol 67(suppl):20-23. 1948.
- Hirsch O. Successful closure of cerebrospinal fluid rhinorrhea by endonasal surgery. Arch Otolaryngol 56:1-13. 1952.
- Kern EB. Pearson BW. McDonald TJ. et al. The transseptal approach to lesions of the pituitary and parasellar regions. Laryngoscope 89:1-34, 1979.
- Vrabec DP. Hallberg OE. Cerebrospinal fluid rhinorrhea by endonasal surgery. Arch Otolaryngol 80:218-229. 1964.
- Wigand ME. Transnasal ethmoidectomy under endoscopical control. Rhinology 19:7-15. 1981.
- Stammberger H. Greistorfer K, Wolf G. et al. Operatives Verschluss von Liquorfisteln der vorderen Schaedelbasis unter inlrathekaler Natriumfluoreszeinanwcndung. Laryngol Rhinol Otol 76:595-607. 1997.
- Lanza D, Douglas A, O'Brien A, et al. Endoscopic repair of cerebrospinal fluid fistulae and encephaloceles. Laryngoscope 106:1119-1125, 1996.
- I I. Gjuric M. Goode U. Keimer H, et al. Endonasal endoscopic closure of cerebrospinal fluid fistulas at the anterior cranial base. Ann Otol Rhinol Laryngol 105:620-623, 1996.
- Visot A, Cophignon J, Derome PJ. Rhinorrhee cerebro-spinale, Sem Hop Paris 3-4:158-166, 1980.
- Dodson DD, Gross CW, Swerdloff JL, Gustafson LM. Transnasal endoscopic repair of cerebrospinal fluid rhinorrhea and skull base defects: A review of twenty-nine cases. Otolaryngol Head Neck Surg 111:600-605. 1994.
- Kirchner FR. Proud GO. Method for identification and localization of cerebrospinal fluid rhinorrhea and otorrhea. Laryngoscope 70:921-931. 1960.
- Wolf G. Greistorfer H. Stammberger H. Der endoskopische Nachweis von Liquorfisteln mittels der Iluoreszeintechnik, Laryngol Rhinol Otol 76. 588-594. 1997.
- Hubbard JL. McDonald TJ, Pearson BW, et al. Spontaneous cerebrospinal fluid rhinorrhea: Evolving concepts in diagnosis and surgical management based on the Mayo Clinic experience from 1970 through 1981. Neurosurgery 16:314-321. 1985.
- Yessenow RS. McCabe B. The osteo-mucoperiosteal flap in repair of cerebrospinal fluid rhinorrhea: A 20-year experience. Otolaryngol Head Neck Surg 101:555-558, 1989.
- Nandapalan V. Watson II). Swift AC. Beta-2-transferrin and cerebrospinal fluid rhinorrhoea. Clin Otolaryngol 21:259-264, 1996.
- Loew F, Pertuiset B. Chaumier FE. et al. Traumatic, spontaneous, and postoperative CSF rhinorrhea. Adv Tech Stand Neurosurg 33:312. 1970.
- Laws RI.. Kern EB. Complications of trans-sphenoidal surgery. Clinical Neurosurgery 23:401-416. 1976.
- Ommaya AK , Di Chiro G, Baldwin M, et al. Non-traumatic cerebrospinal fluid rhinorrhoea. J Neurol Neurosurg Psychiat 31, 214-225, 1968.
- Applebaum EL, Chow JM. Cerebrospinal fluid leaks. Ch. 63 In: Otolaryngology Head & Neck Surgery. 3rd ed. Cummings CW (Ed). Mosby, 1998. 1189-1198.
- Nasseri SS. McCaffrey TV, Kasperhauer JL, et al. A combined, minimally invasive transnasal approach to the sella turcica. Am J Rhinol 12(6):409-416, 1998.
- Hosemann W, Goede U, Sauer M. Wound healing of mucosal autografts for frontal cerebrospinal fluid leaks-clinical and experimental investigations. Rhinology (in press).
- Hosemann W. Nitsche N, Rettinger G. et al. Die endonasale, endoskopisch kontrollierte Versorgung von Duradefekten der Rhinobasis, Laryngol Rhinol Otol 70, 115-119, 1991.
- Nishihira S, McCaffrey T. The use of fibrin glue for the repair of experimental CSF rhinorrhea. Laryngoscope 98:625-627, 1988.
- Moore EJ, Reder PA, Kern EB. Atrophic rhinitis: A review of 222 cases. The case against extensive turbinectomy. Submitted for publication to the American Journal of Rhinology, 1997.