Benign Intracranial Hypertension: a Cause of CSF Rhinorrhoea

JYournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:847-849 847

SHORT REPORT J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.7.847 on 1 July 1994. Downloaded from

Benign intracranial hypertension: a cause of CSF rhinorrhoea

D Clark, P Bullock, T Hui, J Firth

Abstract cranial hypertension.2-5 In this paper the Four patients undergoing treatment for mechanisms by which a CSF leak can occur benign intracranial hypertension pre- are discussed, as well as the difficulties that sented with spontaneous CSF rhinor- may be encountered in securing a direct rhoea. The four patients, all women, repair of the CSF fistula. The distinction were aged between 33 and 44 years. They between high and low pressure CSF leaks is had been receiving treatment for benign emphasised because it is usually important to intracranial hypertension for a period treat the underlying cause of the high pres- ranging from eight months to 11 years, sure, before a satisfactory repair of the CSF before developing the CSF leak. There leak can be effected. was no history of previous head injury and there were no congenital anomalies of the floor of the anterior fossa. The site Patients and methods of the CSF fistula was localised to the Four patients who were treated for benign cribriform plate in all four cases. The intracranial hypertension at the Neurosurgical pathophysiology of the CSF rhinorrhoea Department, University Hospital, Nottingham and the surgical management of this between 1986 and 1990 are described (table). group of patients are discussed. The The four women were aged between 33 and authors propose that benign intracranial 44 years, and had been diagnosed as having hypertension should be included in the benign intracranial hypertension, receiving classification ofhigh pressure" CSF leaks. treatment for a period ranging from eight months to 11 years, before developing CSF (7 Neurol Neurosurg Psychiatry 1994;57:847-849) rhinorrhoea. All the patients had complained of the classic symptoms of headache and Neurosurgical Unit, visual disturbance, and were found to have

University Hospital, http://jnnp.bmj.com/ Queen's Medical Ommaya's original classification of high pres- papilloedema at the time of the original pre- Centre, Nottingham, sure CSF leaks did not include benign sentation. Neurological examination was nor- UK D Clark intracranial hypertension.' He described mal. Diagnosis was by exclusion with a P Bullock tumours and untreated hydrocephalus as the normal CSF under elevated pressure. All T Hui most common causes of this condition. He patients had CT and MRI scans as well as J Firth speculated on a third group such as benign cerebral angiography to exclude other disor- Correspondence to: P Bullock, Department of intracranial hypertension with chronically ders.

Neurosurgery, The raised pressure "placing the fragile roof of the The initial medical management was by on October 1, 2021 by guest. Protected copyright. Maudsley Hospital, Denmark Hill, London nose at risk" and leading to rhinorrhoea, but lumbar puncture, weight reduction, diuretics, SE5 8AZ, United Kingdom. he had not witnessed such a case. Since and steroids. Patients referred for surgery had Received 8 September 1992 Ommaya's classification, several authors progressive visual loss or had failed to and in revised form 28 September 1993. describing the surgical repair of CSF leaks improve with medical therapy. Once the CSF Accepted 9 November 1993 have included patients with benign intra- leak had been diagnosed the site of the fistula was identified in each case by CT scan with fine coronal cuts. The fistulas were all discov- Cerebrospinal fluid leaks ered in the region of the cribriform plate and no congenital anomalies of this area were identified. Traumatic Non-traumatic CASE 1 Accidental latrogenic High pressure Normal p3ressure This 33-year-old female with resistant benign leaks lealiks intracranial hypertension developed a CSF I~~~~~~~~~~~~~~~~~~~~~~~~ rhinorrhoea while on the waiting list for a Tumours Hydrocephalus Benign lumboperitoneal shunt. The CSF leak intracranial resolved after insertion of the shunt. Six hypertension months later the CSF rhinorrhoea recurred Figure 1 Causes of CSF leaks. and the shunt was found to be blocked. 848 Clark, Bullock, Hui, Firth

Initial presentation and treatment Investigations J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.7.847 on 1 July 1994. Downloaded from Patient Age CSFpressure no. Sex (years) Symptoms and signs CT scan (cmH20) Initial treatment 1 Female 33 Headaches Normal: small 30 Multiple lumbar punctures Double vision ventricles Diuretics Gross papilloedema Steroids Enlarged blind spots Weight reduction 2 Female 44 Headaches Normal: normal 34 Multiple lumbar punctures Obscuration ventricle size Steroids Papilloedema 3 Female 42 Dizzy spells Normal: normal 34 Multiple lumbar punctures Papilloedema ventricle size Acetazolamide Enlarged blind spots 4 Female 34 Headaches Normal: small 30 Steroids Blurred vision ventricles Multiple lumbar punctures Enlarged blind spots Thecoperitoneal shunt

Revision of the shunt has been partially suc- was accomplished. After the operation the cessful with the patient still experiencing patient was treated with a lumbar CSF drain intermittent CSF rhinorrhoea and occasional which was continued for five days. She has headaches. The patient had no papilloedema remained symptom-free five years after this or visual disturbance on review at one year procedure. follow up, and she has declined any further intervention. CASE 4 A 34-year-old patient had a lumboperitoneal CASE 2 shunt inserted as treatment for her resistant A 44-year-old female had been treated for benign intracranial hypertension. Six years benign intracranial hypertension for 19 after this procedure the patient developed a months before developing CSF rhinorrhoea. CSF rhinorrhoea, and the shunt was found to Primary closure of the CSF fistula by cran- have migrated and not be functioning. She iotomy was attempted but frustrated by subsequently underwent multiple shunt revi- raised intracranial pressure. A lumboperi- sions and the CSF rhinorrhoea resolved; she toneal shunt was inserted producing resolu- then developed an intracranial aerocele (fig tion of both the rhinorrhoea and symptoms 2). At this point a craniotomy was under- and signs of benign intracranial hypertension. taken to repair the site of the fistula directly. However, following the shunt insertion the At a six month followup there had been no patient complained of low pressure headaches further CSF rhinorrhoea and the aerocele had which persisted, and the shunt had to be resolved. occluded. At followup she remained symptom-free with no recurrence of the CSF leak, presumably as a result of spontaneous resolu- Discussion tion of the benign intracranial hypertension. Benign intracranial hypertension most com- monly affects women of child bearing age and http://jnnp.bmj.com/ CASE 3 is usually a self limiting condition. Most This 42-year-old female had been treated for patients are managed successfully by non-sur- benign intracranial hypertension for more gical methods. A shunting procedure or optic than four years before developing CSF rhin- nerve decompression may be required to con- orrhoea. Primary repair of the CSF fistula trol papilloedema in refractory cases.67 The was undertaken by a craniotomy using an association between benign intracranial extradural approach. At operation, the hypertension and spontaneous CSF rhinor-

intracranial pressure was high but the repair rhoea is rare. on October 1, 2021 by guest. Protected copyright. These four female patients illustrate the more severe form of benign intracranial hypertension. They had all received maxi- Figure 2 Intracranial mum medical treatment and two had aerocoele post lumboperitoneal shunt. described symptoms for over four years. There were no other unusual or unique fea- tures which could help to identify other patients with benign intracranial hypertension at risk of developing a CSF leak. It appears that the raised intracranial pressure was the only aggravating condition. Trauma and congenital anomalies had been excluded. The actual mechanism by which benign intracranial hypertension may precipi- tate a CSF leak is unclear. Locke8 proposed that the exaggerated CSF pulsatile flow, seen with all forms of raised intracranial pressure, could lead to expansion and eventual rupture of the arachnoid sleeve surrounding Benign intracranial hypertension 849

the olfactory filaments which pass through the fistula may not close spontaneously after the pits in the cribriform plate. the shunting procedure. It could also be speculated that the empty In practice direct repair of the CSF fistula J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.7.847 on 1 July 1994. Downloaded from sella syndrome, which has been reported in was found to be difficult, when no shunting association with benign intracranial hyperten- procedure had previously been carried out. sion, may allow the transmission of the Two of the four patients who underwent increased CSF pulsation through the incom- craniotomy as the initial procedure had the petent diaphragma sellae, thereby eroding the intracranial pressure reported as "bursting" at sella and eventually leading to a CSF leak operation, in spite of determined hyperventi- from this site.39 lation and other measures. One of these oper- Six other patients with benign intracranial ations was abandoned before the repair could hypertension who developed spontaneous be effected; the other was successfully com- CSF leaks have previously been included in pleted following the insertion of an external surgical reports on the treatment of non-trau- lumbar CSF drain. Once the intracranial matic CSF rhinorrhoea.2-5 These six patients pressure has been controlled by inserting a have a similar age distribution, sex, and mode lumboperitoneal shunt, the site of the fistula of presentation to the four patients described can be approached directly and a repair here. The cribriform plate was the site of undertaken, using either intracranial or CSF leak in five of the patients and Bejerre et extracranial approaches. Extracranial tech- al3 described one patient who developed a niques have, however, been more recently leak from the sella turcica. From these two described for repair of both the sella and crib- groups of patients, it would appear that the riform regions, and are reported to carry a cribriform plate and the sella region are the reduced morbidity in comparison with the most common sites for the development of a intracranial repair." high pressure CSF leak. In conclusion CSF rhinorrhoea is a rare In this study the site of the CSF fistula was complication of benign intracranial hyperten- adequately demonstrated by coronal CT sion. The cribriform plate, followed by the scans in all four patients, but if this investiga- sella turcica, are the most frequent sites of tion is unrewarding, evaluation by metriza- CSF fistula formation. The shunting proce- mide CT cisternography or CSF radioactive dure may control the benign intracranial isotope tracer studies has been recom- hypertension and the CSF rhinorrhoea, but mended.4 10 to exclude the risk of later meningitis and It is of interest that, during this study, no pneumocephalus, a direct repair of the fistula patients' vision deteriorated after develop- is also recommended. ment of the CSF leak, but high pressure CSF rhinorrhoea rarely stops of its own accord and places the patient at risk of meningitis. Surgical repair was therefore recommended 1 Ommaya AK. Cerebrospinal fluid rhinorrhoea. Neurology 1964;27:214-6. to prevent this complication. 2 Brisman R, Hughes JEO, Mount LA. Cerebrospinal fluid It is usually important to treat the underly- rhinorrhoea. Arch Neurol 1970;22:245-52. 3 Bjerre P, Lindholm J, Glydensted C. Pathogenesis of non- ing cause of the high pressure before a satis- traumatic cerebrospinal fluid rhinorrhoea. Acta Neurol factory repair of the CSF leak can be Scand 1982;66:180-92. 4 Eljamel MSM, Foy PM. Non-traumatic CSF fistulae: undertaken. A lumboperitoneal shunt has clinical history and management. Br Jf Neurosurg 1991; http://jnnp.bmj.com/ proved useful in controlling the raised 5:275-9. 5 Tolley NS, Brookes GB. Surgical management of cere- intracranial pressure in benign intracranial brospinal fluid rhinorrhoea. Jf R Coll Surg Edinb 1992; hypertension. However, the fistula may not 37:12-5. 6 Johnson I, Besser M, Morgan MK. Cerebrospinal fluid heal completely after a shunting procedure diversion in the treatment of benign intracranial hyper- alone, and the patient may be exposed to an tension. J7 Neurosurg 1988;69: 195-202. 7 Kelman SE, Heaps R, Wolf A, Elman MJ. Optic nerve increased risk of infection because the gradi- decompression surgery improves visual function in ent across the fistula has been reversed. patients with pseudotumour cerebri. Neurosurgery 1992;

30:391-5. on October 1, 2021 by guest. Protected copyright. Eljamel and Foy4 reported this complication 8 Locke CE. The spontaneous escape of cerebrospinal fluid in a patient who developed meningitis after through the nose. Arch Neurol Psychiatry (Chicago) 1926;15:309-24. receiving a lumboperitoneal shunt for CSF 9 Weisberg LA, Housepian EM, Saur DP. Empty sella syn- rhinorrhoea. Indeed, two patients in this drome as complication of benign intracranial hypertension. Jf Neurosurg 1977;43:177-80. study presented with a recurrent CSF rhinor- 10 Ahmade J, Weiss MH, Segall HF, Schultz DH, Lee CS, rhoea as a result of shunt failure. One patient Giennotta SL. Evaluation of cerebrospinal fluid rhinorrhoea by Metrizamide computed tomographic cistern- (case 4) had an acute exacerbation of her ography. Neurosurgery 1985;16:54-60. headaches after blowing her nose and was 11 Hubbard JL, McDonald TJ, Pearson BW, Laws ER. Spontaneous cerebrospinal fluid rhinorrhoea. Evolving discovered to have an extensive pneumo- concepts in diagnosis and surgical management from cephalus (fig 2). This is further evidence that 1970 through 1981. Neurosurgery 1985;16:314-21.