- Email: [email protected]
Endoscopic transnasal intradural repair of anterior skull base cerebrospinal fluid fistulae
Journal of Clinical Neuroscience (2004) 11(6), 597–599 0967-5868/$ - see front matter ª 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jocn.2003.09.011
Clinical study
Endoscopic transnasal intradural repair of anterior skull base cerebrospinal fluid fistulae R.J.S. Briggs1 1 2
FRACS FACS,
P.J. Wormald2 MD FRACS FRCS FCS(SA)
Department of Surgery, Royal Melbourne Hospital, Department of Otolaryngology, University of Melbourne, Melbourne, Vic., Australia, Department of Surgery – Otolaryngology, Head and Neck Surgery, Adelaide and Flinders Universities of Australia, Adelaide, Australia
Summary This study presents the techniques and results of endoscopic diagnosis and repair of cerebrospinal fluid (CSF) fistulae involving the anterior skull base and paranasal sinuses. Design. A retrospective cohort study of all consecutive patients undergoing endoscopic repair of anterior skull base CSF fistulae. Setting. Tertiary referral institutions. Materials and methods. Fifty-two patients underwent endoscopic repair of CSF fistula. Thirteen cases were traumatic in origin, 11 spontaneous not associated with meningoencephalocele and 12 with meningoencephalocele. Eleven were iatrogenic and five associated with transphenoidal pituitary surgery, two acute and three delayed following radiotherapy. The average age of patients was 43 and the male to female ratio was 2:1. A variety of techniques were used to repair the dural defect. In the majority of cases placement of a fat plug on the intra-cranial surface of the dura was performed. Results. Forty-seven of the 52 patients had successful primary endoscopic repair of the CSF fistula and skull base defect. Five patients required a repeat procedure due to early failure of the repair. After an average follow-up of 27 months no patient has had any recurrence of leak giving a primary closure success rate of 90% and secondary closure rate of 100%. Conclusions. The endoscopic transnasal approach for repair of anterior skull base CSF fistula is a reliable technique and is now the procedure of choice for patients presenting with this problem. ª 2004 Elsevier Ltd. All rights reserved.
INTRODUCTION The transnasal endoscopic approach is now recognised as the preferred method for the diagnosis and repair of anterior skull base cerebrospinal fluid leaks.1–4 The endoscopic transnasal approach carries minimal morbidity, has been demonstrated to be effective and has replaced transcranial techniques of repair as the first line procedure for these problems. Anterior skull base cerebrospinal fluid fistula may occur following anterior skull base fractures, surgery on the paranasal sinuses or be spontaneous. Spontaneous cerebrospinal fluid (CSF) leaks may occur in the cribriform plate region or in the sphenoid. They may also be associated with a congenital meningoencephalocele which occur typically in the frontoethmoid or sphenoid sinus. Usually, the cause is uncertain but there has recently been an association between middle aged overweight female patients and spontaneous CSF leaks.5 The cause of this is thought to be a malabsorption of CSF from dilated arachnoid granulations. Usually, these granulations are associated with large intracranial venous sinuses and the CSF drains from them into the venous system. However, in the very pneumatized sphenoid these granulations may not be associated with a venous sinus. Chronic coughing and CSF pulsation is thought to erode the thin bone of the skull base in this region with a meningocoele and finally a CSF leak developing. There may also be an association with benign intracranial hypertension increasing CSF pressure. However, in most cases the reason for patients developing a spontaneous CSF leak is uncertain. In some patients, the CSF leak may be intermittent and this creates a
Received 8 July 2003 Accepted 2 September 2003 Correspondence to: Robert J.S. Briggs FRACS, FACS, Department of Otolaryngology, University of Melbourne, 32 Gisborne Street, East Melbourne, Vic. 3002, Australia. Tel.: +61-3-9416-4133; Fax: +61-3-9416-1343; E-mail: [email protected]
problem both with confirming that the fluid is CSF and also with localisation of the site of leak. The use of Beta-2-asialotransferrin testing of the fluid is very specific for CSF and the most effective test for CSF. Previously used investigations such as nuclear medicine studies using intrathecal isotope have a very poor anatomic specificity and high false positive rate and should no longer be used in the investigation of these patients. Although there have been a large number of techniques described for the transnasal endoscopic closure of CSF leaks, there are few large series where predominantly one technique has been used. The authors have a large experience with the fat plug technique where a fat plug is placed through the defect onto the intra-cranial surface of the dura and reinforced with a mucoperiosteal graft on the nasal surface.6 This technique has been termed the “bath plug” technique as it uses the CSF water pressure to help maintain the seal of the fat on the intracranial surface of the defect much in the same way that intracranial techniques lay dura or fascia lata over the defect on the intracranial surface of the anterior skull base. In addition a free mucosal graft reinforces the closure on the nasal surface of the graft. We present the results of endoscopic management of a large series of patients with CSF fistula using both this and other techniques. MATERIALS AND METHODS All consecutive patients with anterior skull base cerebrospinal fluid fistulae managed by the authors between 1995 and 2002 were included in this study. In a number of cases cerebrospinal fluid rhinorrhoea was due to a CSF fistula into the temporal bone with leakage down the Eustachian tube and these patients were not included. There were 52 patients with 35 males and 17 females and a male to female ratio of 2:1. The average age was 43 years (standard deviation ¼ 20.4 years). Eleven patients had iatrogenic CSF leaks, 11 had spontaneous leaks (not associated with meningocele) and 12 patients had meningoceles or meningoencephaloceles, 13 had 597
598 Briggs and Wormald
post-traumatic leaks and five had CSF leaks associated with prior pituitary fossa surgery, two perioperatively following transphenoidal pituitary surgery and three being very delayed following both transphenoidal pituitary surgery and radiation therapy. Three patients in the series had had previous transnasal attempts to repair a sphenoid sinus leak by fat obliteration of the sphenoid at another institution and two patients with anterior skull base fractures had undergone prior transcranial repairs which had been unsuccessful in controlling the CSF fistula. No other patients had undergone prior surgery for CSF leak closure. Four patients had gross pneumocephalus with obtunded mental status in association with their CSF fistula. In the majority of cases (90%) the “bath plug” technique of intracranial placement of a fat graft was used as the method of repair. In five cases a free mucosal graft was used and in one case a local rotation mucosal flap. In several cases where larger defects were involved a bone graft was placed in addition to the “bath plug” fat graft and the mucoperiosteal graft. SURGICAL TECHNIQUE In all cases intrathecal Fluorescein is used to stain the CSF. This is placed via a lumbar drain before or after induction of general anaesthesia and is used to identify the site of the fistula. The fistula cannot be closed if the site cannot be identified. In most cases a fine slice high definition CT scan and T2-weighted MRI scan will be sufficient to identify the site of the leak but in some patients, specifically patients with small intermittent leaks, identification can be difficult. The Fluorescein is also helpful at completion of the repair to ensure that a watertight seal has been achieved. About 0.2–0.5 ml (depending upon patients weight) of 5% Fluorescein is mixed with 10 ml of either CSF or Ringers solution. A filter is attached to the lumbar drain and the Fluorescein stained fluid is reinjected into the intrathecal space at the rate of 1 ml/min. The patient is asked to identify any paraesthesia during the injection and the injection is discontinued if this occurs. However, no such complication was seen in this series of patients. The other possible complication is grand mal seizures but again this was not seen in any patients in this series. The patient is then given a general anaesthetic and positioned head down. Intravenous antibiotic prophylaxis is administered and the nasal mucosa is prepared with a topical decongestant. If there is free flow of CSF the Fluorescein colour is clearly visible pooling in the nasopharynx on the side of the fistula on endoscopic examination. The trail of Fluorescein can then be followed to the region of the leak. In some cases the Fluorescein may not be visible initially. If there is doubt about the site or presence of a CSF leak the patient should be positioned head down and a forced inspiration (Valsalva) manoeuvre performed by the anaesthetist to increase the CSF pressure. This manoeuvre should be repeated a number of times and if the leak can still not be identified the patient should be placed in the head down position for a further 30 min. If after this period there is still no Fluorescein visible then a blue light filter should be used. The blue light allows extremely small quantities of Fluorescein to be identified. If there is still no Fluorescein visible then the CSF pressure can be artificially raised by the injection of Ringers lactate solution into the CSF space through the lumbar spinal drain. This may be done in 20 ml aliquots with the CSF pressure being checked after each aliquot to ensure rapid increase does not occur, and the forced inspiration manoeuvre repeated. A total of between 100 and 150 ml of Ringers solution may be added in this way. Once the fistula is identified the mucosa around the defect is removed with a sickle knife exposing the bone around the skull Journal of Clinical Neuroscience (2004) 11(6), 597–599
base defect for at least a radius of 5 mm. Any prolapsing dura is removed and the size of defect is measured. If the defect is less than 1 mm or there is concern about proximity of intracranial structures such as the optic nerve or vessels, it may be preferable to avoid the fat plug and only place a mucosal graft over the defect. In most cases (47 out of 52) the authors have preferred the “bath plug” technique for closure. Where there is a small dural defect and the “bath plug” technique is used it may be necessary to gently enlarge the defect with a probe or curette to allow subsequent insertion of the fat graft through the defect. Ear lobe fat is used because of its rather fibrous consistency unless the defect is very large in which case abdominal fat, preferably including Scarpas fascia, can be used. The fat graft is prepared with the diameter the same size as the defect and about 1–1.5 cm long. A 4.0 Vicryl suture is secured at one end of the graft and passed longitudinally through the graft. The fat plug is gently manipulated through the defect. Care is taken to pass a little fat through the defect with each manipulation of the probe as this limits the distance that the probe needs to be introduced intracranially and allows the fat to gently push away any adjacent intracranial structures as the fat enters the intracranial cavity. This significantly limits any possible damage to adjacent intracranial vascular or neural structures. The fat plug is supported with a probe while gentle traction is placed on the suture which causes the fat graft to concertina and expand in width hence securely seals the dural defect from above. The repair is then reinforced with a free mucoperiosteal graft on the nasal side. For defects greater than 1 cm, it may be preferable to position a bone graft between the fat graft and mucoperiosteal graft. Bone may be harvested from the vomer and also shaped so that the length is slightly longer than the diameter of the defect whilst the width is slightly narrower so that it may be inserted through the defect and positioned to just overlap the bony edges and hence provide rigid support for the fat graft above. A hole is made in the centre of the bone graft to allow it to be slid up along the suture. If the closure is effective fibrin glue is used to help seal the fat graft and mucoperiosteal graft in position. The suture may be either cut or sutured to the septum to provide some continued traction. Immediately following the repair, a forced inspiration manoeuvre is again performed by the anaesthetist and the seal is checked. Fluorescein should not be visible during the forced inspiration. If the repair appears secure a dissolving dressing such as gelatin sponge is applied. When the method of repair used is only a mucoperiosteal graft on the nasal surface of the defect without the intra-cranial fat plug, it is more important that the graft is held securely in position from below so that the seal remains effective. In this situation the authors have used an initial layer of dressing such as gelatin sponge and then reinforced this with a more solid removable pack such as ribbon gauze impregnated with Bismuth Iodoform Paraffin paste (BIPP). At completion of the procedure the lumbar spinal drain may be removed, or is left in place for 24 h and positioned at the shoulder of the patient to ensure that CSF is only drained if the intracranial pressure is raised if, for example, the patient coughs or strains during recovery. The patient is then mobilized and discharged the following day. RESULTS Of the 52 patients who underwent endoscopic repair of their CSF fistula and skull base defect, 47 have had successful primary closure with no recurrence of the CSF leak. Five patients required a repeat procedure due to early failure of the repair. After an average follow-up of 27 months no patient has had any recurrence of leak giving a primary closure success rate of 90% and secondary closure success rate of 100%. Of the five patients requiring ª 2004 Elsevier Ltd. All rights reserved.
Cerebrospinal fluid fistulae 599
a second procedure, two were due to early post-operative recurrence of CSF leak associated with displacement of the fat plug graft. The other three had recurrence of leak between four and six weeks post-operatively. Two were patients with previous transphenoidal pituitary surgery. In both these cases a free mucoperiosteal graft was placed over the defect and the graft held in place by further packing of the sphenoid with abdominal fat. Presumably, the fat within the sphenoid did not provide adequate support for the graft. In each case at revision surgery, a repeat free mucoperiosteal graft was used but this was supported by BIPP gauze within the sphenoid and secure healing achieved without recurrence. The fifth patient to require a revision procedure had a laterally placed sphenoid meningocele requiring an extensive endoscopic procedure in order to obtain sufficient access to repair. A recurrent leak at one month was successfully repaired using a “bath plug” procedure. There were no complications associated with the intrathecal Fluorescein use specifically no paraesthesias and no seizures. There were no complications associated with introduction of the fat plug into the intra-cranial cavity. DISCUSSION It is clear that endoscopic repair of CSF fistula is effective in the hands of experienced surgeons. A number of authors report high success rates with successful primary repair in 90% or more of cases.7–9 Potential benefits of an endoscopic transnasal approach rather than a transcranial approach is the minimally invasive procedure and a high success rate compared to the average success rate of 70% achieved with transcranial approach.10 Achieving an effective repair of CSF fistula via a transcranial approach can be difficult and may be unsuccessful as the surgeon will in some patients not be able to accurately identify the site of the leak. The avoidance of a craniotomy and the associated risks of a neurosurgical procedure is clearly an advantage to the patient. Transcranial approaches for repair of anterior skull based defect in the region of the paranasal sinuses usually result in destruction of olfactory function either unilateral or often bilateral with permanent anosmia. Relative to a craniotomy, the endoscopic transnasal approach is a minimally invasive procedure often requiring only a one or two night inpatient stay with little associated morbidity. The progressive development of endoscopic techniques used for paranasal sinus surgery have facilitated a minimally invasive and effective CSF fistula repair and have facilitated techniques for the diagnosis and localisation of CSF fistulae. Confirmation of CSF rhinorrhoea can be confidently achieved with testing of fluid for Beta-2-asialotransferrin. In most cases imaging by high-resolution CT in the coronal plane in combination with T2-weighted magnetic resonance imaging targeting the olfactory fossa and the paranasal sinus region, usually provides sufficient information to identify the site of CSF fistulae. The value of intrathecal Fluorescein is that it provides a highly sensitive mechanism to identify the site of the leak. In addition, once the leak has been closed the seal can be tested by a forced inspiration manoeuvre. Even the smallest amount of Fluorescein stained CSF can be seen leaking from the dural seal. If this is seen the repair can be re-done. Fluorescein instilled into the subarachnoid space does have the potential to cause complications such as peripheral paraesthesias and seizures.11 However, with the small doses currently recommended there have been large series of patients without any complications.6;12;13 Once the site of CSF fistula is identified a variety of techniques are available for the repair of the defect using an endoscopic approach. A mucoperiosteal graft can be applied to the intranasal surface of the bone of the skull base, either as a free graft or as a ª 2004 Elsevier Ltd. All rights reserved.
pedicled local rotation flap. This can be very effective for small defects (less than 3 mm) but for larger defects maintaining a watertight seal and hence effective repair can be difficult. Some authors have described underlay graft techniques where a fascia or cartilage graft is placed between the skull base and dura allowing the CSF pressure to push the dura against the graft. This may not be an easy dissection, however, particularly adjacent to the cribriform plate. Although the “bath plug” technique of repair was utilised in the majority of cases in this series, other authors report equally high success rates with onlay mucoperiosteal grafts without any intradural component. The authors believe that the “bath plug” technique is very suitable for repair of CSF fistulae, because it does provide an intradural seal. The hydrostatic pressure from CSF may well increase the “bath plug” effect by pushing the fat graft into the defect. For inferiorly placed onlay type extradural grafts, the graft must be securely packed against the hydrostatic pressure whilst healing occurs otherwise the repair fails. The disadvantage of the “bath plug” technique is that it does require some minimal intradural dissection which carries a potential risk to adjacent intracranial structures and potentially can introduce of infection. The operating surgeon must be aware of intracranial neural or vascular structures adjacent to the site of fistula and if there is any risk of injury this may influence the type of repair used. For example, for a lateral sphenoid defect adjacent to the optic nerve or internal carotid artery, it is safer to use a graft on the nasal surface only rather than attempt to introduce a “bath plug”. CONCLUSION The endoscopic transnasal technique for repair of anterior skull base CSF fistula is safe and effective and is now the procedure of choice for patients presenting with CSF rhinorrhoea due to anterior skull base defects. This series confirms that the “bath plug” technique of endoscopic repair is effective and reliable and may be used for a variety of types of defects. REFERENCES 1. Marshall AJ, Jones NS et al. CSF rhinorrhoea: the place of endoscopic sinus surgery. British Journal of Neurosurgery 2001; 15(1): 8–12. 2. Hughes RGM, Jones NS et al. The endoscopic treatment of cerebrospinal fluid rhinorrhoea: the Nottingham experience. The Journal of Laryngology and Otology 1997; 111(February): 125–128. 3. Burns JA, Dodson EE, Gross CW. Transnasal endoscopic repair of cranionasal fistulae: a refined technique with long-term follow-up. Laryngoscope 1996; 106: 1080–1083. 4. Hegazy H, Carrau R et al. Transnasal endoscopic repair of cerebrospinal fluid rhinorrhea: a meta-analysis. Laryngoscope 2000; 110(7): 1166–1172. 5. Badia L, Loughran S et al. Primary spontaneous cerebrospinal fluid rhinorrhoea and obesity. American Journal of Rhinology 2001; 15: 117–119. 6. Wormald PJ, McDonogh M. ‘Bath-plug’ technique for the endoscopic management of cerebrospinal fluid leaks. The Journal of Laryngology and Otology 1997; 111(November): 1042–1046. 7. Zweig JL, Carrau RL et al. Endoscopic repair of cerebrospinal fluid leaks to the sinonasal tract: predictors of success. Otolaryngology – Head and Neck Surgery 2000; 123(3): 195–201. 8. Lanza DC, O’Brien DA, Kennedy DW. Endoscopic repair of cerebrospinal fluid fistulae and encephaloceles. Laryngoscope 1996; 106: 1119–1125. 9. Chin G, Rice DH. Transnasal endoscopic closure of cerebrospinal fluid leaks. Laryngoscope 2003; 113: 136–138. 10. Aarabi B, Leibrock LG. Neurosurgical approaches to cerebrospinal fluid rhinorrhea. Ear, Nose and Throat Journal 1992; 71(7): 300–305. 11. Moseley JI, Carton CA et al. Spectrum of complications in the use of intrathecal fluorescein. Journal of Neurosurgery 1978; 48(May): 765–767. 12. Mattox DE, Kennedy DW. Endoscopic management of cerebrospinal fluid leaks and cephaloceles. Laryngoscope 1990; 100: 857–862. 13. Casiano RR, Jassir D. Endoscopic cerebrospinal fluid rhinorrhea repair: is a lumbar drain necessary? Otolaryngology – Head and Neck Surgery 1999; 121: 745–750.
Journal of Clinical Neuroscience (2004) 11(6), 597–599
Clinical study
Endoscopic transnasal intradural repair of anterior skull base cerebrospinal fluid fistulae R.J.S. Briggs1 1 2
FRACS FACS,
P.J. Wormald2 MD FRACS FRCS FCS(SA)
Department of Surgery, Royal Melbourne Hospital, Department of Otolaryngology, University of Melbourne, Melbourne, Vic., Australia, Department of Surgery – Otolaryngology, Head and Neck Surgery, Adelaide and Flinders Universities of Australia, Adelaide, Australia
Summary This study presents the techniques and results of endoscopic diagnosis and repair of cerebrospinal fluid (CSF) fistulae involving the anterior skull base and paranasal sinuses. Design. A retrospective cohort study of all consecutive patients undergoing endoscopic repair of anterior skull base CSF fistulae. Setting. Tertiary referral institutions. Materials and methods. Fifty-two patients underwent endoscopic repair of CSF fistula. Thirteen cases were traumatic in origin, 11 spontaneous not associated with meningoencephalocele and 12 with meningoencephalocele. Eleven were iatrogenic and five associated with transphenoidal pituitary surgery, two acute and three delayed following radiotherapy. The average age of patients was 43 and the male to female ratio was 2:1. A variety of techniques were used to repair the dural defect. In the majority of cases placement of a fat plug on the intra-cranial surface of the dura was performed. Results. Forty-seven of the 52 patients had successful primary endoscopic repair of the CSF fistula and skull base defect. Five patients required a repeat procedure due to early failure of the repair. After an average follow-up of 27 months no patient has had any recurrence of leak giving a primary closure success rate of 90% and secondary closure rate of 100%. Conclusions. The endoscopic transnasal approach for repair of anterior skull base CSF fistula is a reliable technique and is now the procedure of choice for patients presenting with this problem. ª 2004 Elsevier Ltd. All rights reserved.
INTRODUCTION The transnasal endoscopic approach is now recognised as the preferred method for the diagnosis and repair of anterior skull base cerebrospinal fluid leaks.1–4 The endoscopic transnasal approach carries minimal morbidity, has been demonstrated to be effective and has replaced transcranial techniques of repair as the first line procedure for these problems. Anterior skull base cerebrospinal fluid fistula may occur following anterior skull base fractures, surgery on the paranasal sinuses or be spontaneous. Spontaneous cerebrospinal fluid (CSF) leaks may occur in the cribriform plate region or in the sphenoid. They may also be associated with a congenital meningoencephalocele which occur typically in the frontoethmoid or sphenoid sinus. Usually, the cause is uncertain but there has recently been an association between middle aged overweight female patients and spontaneous CSF leaks.5 The cause of this is thought to be a malabsorption of CSF from dilated arachnoid granulations. Usually, these granulations are associated with large intracranial venous sinuses and the CSF drains from them into the venous system. However, in the very pneumatized sphenoid these granulations may not be associated with a venous sinus. Chronic coughing and CSF pulsation is thought to erode the thin bone of the skull base in this region with a meningocoele and finally a CSF leak developing. There may also be an association with benign intracranial hypertension increasing CSF pressure. However, in most cases the reason for patients developing a spontaneous CSF leak is uncertain. In some patients, the CSF leak may be intermittent and this creates a
Received 8 July 2003 Accepted 2 September 2003 Correspondence to: Robert J.S. Briggs FRACS, FACS, Department of Otolaryngology, University of Melbourne, 32 Gisborne Street, East Melbourne, Vic. 3002, Australia. Tel.: +61-3-9416-4133; Fax: +61-3-9416-1343; E-mail: [email protected]
problem both with confirming that the fluid is CSF and also with localisation of the site of leak. The use of Beta-2-asialotransferrin testing of the fluid is very specific for CSF and the most effective test for CSF. Previously used investigations such as nuclear medicine studies using intrathecal isotope have a very poor anatomic specificity and high false positive rate and should no longer be used in the investigation of these patients. Although there have been a large number of techniques described for the transnasal endoscopic closure of CSF leaks, there are few large series where predominantly one technique has been used. The authors have a large experience with the fat plug technique where a fat plug is placed through the defect onto the intra-cranial surface of the dura and reinforced with a mucoperiosteal graft on the nasal surface.6 This technique has been termed the “bath plug” technique as it uses the CSF water pressure to help maintain the seal of the fat on the intracranial surface of the defect much in the same way that intracranial techniques lay dura or fascia lata over the defect on the intracranial surface of the anterior skull base. In addition a free mucosal graft reinforces the closure on the nasal surface of the graft. We present the results of endoscopic management of a large series of patients with CSF fistula using both this and other techniques. MATERIALS AND METHODS All consecutive patients with anterior skull base cerebrospinal fluid fistulae managed by the authors between 1995 and 2002 were included in this study. In a number of cases cerebrospinal fluid rhinorrhoea was due to a CSF fistula into the temporal bone with leakage down the Eustachian tube and these patients were not included. There were 52 patients with 35 males and 17 females and a male to female ratio of 2:1. The average age was 43 years (standard deviation ¼ 20.4 years). Eleven patients had iatrogenic CSF leaks, 11 had spontaneous leaks (not associated with meningocele) and 12 patients had meningoceles or meningoencephaloceles, 13 had 597
598 Briggs and Wormald
post-traumatic leaks and five had CSF leaks associated with prior pituitary fossa surgery, two perioperatively following transphenoidal pituitary surgery and three being very delayed following both transphenoidal pituitary surgery and radiation therapy. Three patients in the series had had previous transnasal attempts to repair a sphenoid sinus leak by fat obliteration of the sphenoid at another institution and two patients with anterior skull base fractures had undergone prior transcranial repairs which had been unsuccessful in controlling the CSF fistula. No other patients had undergone prior surgery for CSF leak closure. Four patients had gross pneumocephalus with obtunded mental status in association with their CSF fistula. In the majority of cases (90%) the “bath plug” technique of intracranial placement of a fat graft was used as the method of repair. In five cases a free mucosal graft was used and in one case a local rotation mucosal flap. In several cases where larger defects were involved a bone graft was placed in addition to the “bath plug” fat graft and the mucoperiosteal graft. SURGICAL TECHNIQUE In all cases intrathecal Fluorescein is used to stain the CSF. This is placed via a lumbar drain before or after induction of general anaesthesia and is used to identify the site of the fistula. The fistula cannot be closed if the site cannot be identified. In most cases a fine slice high definition CT scan and T2-weighted MRI scan will be sufficient to identify the site of the leak but in some patients, specifically patients with small intermittent leaks, identification can be difficult. The Fluorescein is also helpful at completion of the repair to ensure that a watertight seal has been achieved. About 0.2–0.5 ml (depending upon patients weight) of 5% Fluorescein is mixed with 10 ml of either CSF or Ringers solution. A filter is attached to the lumbar drain and the Fluorescein stained fluid is reinjected into the intrathecal space at the rate of 1 ml/min. The patient is asked to identify any paraesthesia during the injection and the injection is discontinued if this occurs. However, no such complication was seen in this series of patients. The other possible complication is grand mal seizures but again this was not seen in any patients in this series. The patient is then given a general anaesthetic and positioned head down. Intravenous antibiotic prophylaxis is administered and the nasal mucosa is prepared with a topical decongestant. If there is free flow of CSF the Fluorescein colour is clearly visible pooling in the nasopharynx on the side of the fistula on endoscopic examination. The trail of Fluorescein can then be followed to the region of the leak. In some cases the Fluorescein may not be visible initially. If there is doubt about the site or presence of a CSF leak the patient should be positioned head down and a forced inspiration (Valsalva) manoeuvre performed by the anaesthetist to increase the CSF pressure. This manoeuvre should be repeated a number of times and if the leak can still not be identified the patient should be placed in the head down position for a further 30 min. If after this period there is still no Fluorescein visible then a blue light filter should be used. The blue light allows extremely small quantities of Fluorescein to be identified. If there is still no Fluorescein visible then the CSF pressure can be artificially raised by the injection of Ringers lactate solution into the CSF space through the lumbar spinal drain. This may be done in 20 ml aliquots with the CSF pressure being checked after each aliquot to ensure rapid increase does not occur, and the forced inspiration manoeuvre repeated. A total of between 100 and 150 ml of Ringers solution may be added in this way. Once the fistula is identified the mucosa around the defect is removed with a sickle knife exposing the bone around the skull Journal of Clinical Neuroscience (2004) 11(6), 597–599
base defect for at least a radius of 5 mm. Any prolapsing dura is removed and the size of defect is measured. If the defect is less than 1 mm or there is concern about proximity of intracranial structures such as the optic nerve or vessels, it may be preferable to avoid the fat plug and only place a mucosal graft over the defect. In most cases (47 out of 52) the authors have preferred the “bath plug” technique for closure. Where there is a small dural defect and the “bath plug” technique is used it may be necessary to gently enlarge the defect with a probe or curette to allow subsequent insertion of the fat graft through the defect. Ear lobe fat is used because of its rather fibrous consistency unless the defect is very large in which case abdominal fat, preferably including Scarpas fascia, can be used. The fat graft is prepared with the diameter the same size as the defect and about 1–1.5 cm long. A 4.0 Vicryl suture is secured at one end of the graft and passed longitudinally through the graft. The fat plug is gently manipulated through the defect. Care is taken to pass a little fat through the defect with each manipulation of the probe as this limits the distance that the probe needs to be introduced intracranially and allows the fat to gently push away any adjacent intracranial structures as the fat enters the intracranial cavity. This significantly limits any possible damage to adjacent intracranial vascular or neural structures. The fat plug is supported with a probe while gentle traction is placed on the suture which causes the fat graft to concertina and expand in width hence securely seals the dural defect from above. The repair is then reinforced with a free mucoperiosteal graft on the nasal side. For defects greater than 1 cm, it may be preferable to position a bone graft between the fat graft and mucoperiosteal graft. Bone may be harvested from the vomer and also shaped so that the length is slightly longer than the diameter of the defect whilst the width is slightly narrower so that it may be inserted through the defect and positioned to just overlap the bony edges and hence provide rigid support for the fat graft above. A hole is made in the centre of the bone graft to allow it to be slid up along the suture. If the closure is effective fibrin glue is used to help seal the fat graft and mucoperiosteal graft in position. The suture may be either cut or sutured to the septum to provide some continued traction. Immediately following the repair, a forced inspiration manoeuvre is again performed by the anaesthetist and the seal is checked. Fluorescein should not be visible during the forced inspiration. If the repair appears secure a dissolving dressing such as gelatin sponge is applied. When the method of repair used is only a mucoperiosteal graft on the nasal surface of the defect without the intra-cranial fat plug, it is more important that the graft is held securely in position from below so that the seal remains effective. In this situation the authors have used an initial layer of dressing such as gelatin sponge and then reinforced this with a more solid removable pack such as ribbon gauze impregnated with Bismuth Iodoform Paraffin paste (BIPP). At completion of the procedure the lumbar spinal drain may be removed, or is left in place for 24 h and positioned at the shoulder of the patient to ensure that CSF is only drained if the intracranial pressure is raised if, for example, the patient coughs or strains during recovery. The patient is then mobilized and discharged the following day. RESULTS Of the 52 patients who underwent endoscopic repair of their CSF fistula and skull base defect, 47 have had successful primary closure with no recurrence of the CSF leak. Five patients required a repeat procedure due to early failure of the repair. After an average follow-up of 27 months no patient has had any recurrence of leak giving a primary closure success rate of 90% and secondary closure success rate of 100%. Of the five patients requiring ª 2004 Elsevier Ltd. All rights reserved.
Cerebrospinal fluid fistulae 599
a second procedure, two were due to early post-operative recurrence of CSF leak associated with displacement of the fat plug graft. The other three had recurrence of leak between four and six weeks post-operatively. Two were patients with previous transphenoidal pituitary surgery. In both these cases a free mucoperiosteal graft was placed over the defect and the graft held in place by further packing of the sphenoid with abdominal fat. Presumably, the fat within the sphenoid did not provide adequate support for the graft. In each case at revision surgery, a repeat free mucoperiosteal graft was used but this was supported by BIPP gauze within the sphenoid and secure healing achieved without recurrence. The fifth patient to require a revision procedure had a laterally placed sphenoid meningocele requiring an extensive endoscopic procedure in order to obtain sufficient access to repair. A recurrent leak at one month was successfully repaired using a “bath plug” procedure. There were no complications associated with the intrathecal Fluorescein use specifically no paraesthesias and no seizures. There were no complications associated with introduction of the fat plug into the intra-cranial cavity. DISCUSSION It is clear that endoscopic repair of CSF fistula is effective in the hands of experienced surgeons. A number of authors report high success rates with successful primary repair in 90% or more of cases.7–9 Potential benefits of an endoscopic transnasal approach rather than a transcranial approach is the minimally invasive procedure and a high success rate compared to the average success rate of 70% achieved with transcranial approach.10 Achieving an effective repair of CSF fistula via a transcranial approach can be difficult and may be unsuccessful as the surgeon will in some patients not be able to accurately identify the site of the leak. The avoidance of a craniotomy and the associated risks of a neurosurgical procedure is clearly an advantage to the patient. Transcranial approaches for repair of anterior skull based defect in the region of the paranasal sinuses usually result in destruction of olfactory function either unilateral or often bilateral with permanent anosmia. Relative to a craniotomy, the endoscopic transnasal approach is a minimally invasive procedure often requiring only a one or two night inpatient stay with little associated morbidity. The progressive development of endoscopic techniques used for paranasal sinus surgery have facilitated a minimally invasive and effective CSF fistula repair and have facilitated techniques for the diagnosis and localisation of CSF fistulae. Confirmation of CSF rhinorrhoea can be confidently achieved with testing of fluid for Beta-2-asialotransferrin. In most cases imaging by high-resolution CT in the coronal plane in combination with T2-weighted magnetic resonance imaging targeting the olfactory fossa and the paranasal sinus region, usually provides sufficient information to identify the site of CSF fistulae. The value of intrathecal Fluorescein is that it provides a highly sensitive mechanism to identify the site of the leak. In addition, once the leak has been closed the seal can be tested by a forced inspiration manoeuvre. Even the smallest amount of Fluorescein stained CSF can be seen leaking from the dural seal. If this is seen the repair can be re-done. Fluorescein instilled into the subarachnoid space does have the potential to cause complications such as peripheral paraesthesias and seizures.11 However, with the small doses currently recommended there have been large series of patients without any complications.6;12;13 Once the site of CSF fistula is identified a variety of techniques are available for the repair of the defect using an endoscopic approach. A mucoperiosteal graft can be applied to the intranasal surface of the bone of the skull base, either as a free graft or as a ª 2004 Elsevier Ltd. All rights reserved.
pedicled local rotation flap. This can be very effective for small defects (less than 3 mm) but for larger defects maintaining a watertight seal and hence effective repair can be difficult. Some authors have described underlay graft techniques where a fascia or cartilage graft is placed between the skull base and dura allowing the CSF pressure to push the dura against the graft. This may not be an easy dissection, however, particularly adjacent to the cribriform plate. Although the “bath plug” technique of repair was utilised in the majority of cases in this series, other authors report equally high success rates with onlay mucoperiosteal grafts without any intradural component. The authors believe that the “bath plug” technique is very suitable for repair of CSF fistulae, because it does provide an intradural seal. The hydrostatic pressure from CSF may well increase the “bath plug” effect by pushing the fat graft into the defect. For inferiorly placed onlay type extradural grafts, the graft must be securely packed against the hydrostatic pressure whilst healing occurs otherwise the repair fails. The disadvantage of the “bath plug” technique is that it does require some minimal intradural dissection which carries a potential risk to adjacent intracranial structures and potentially can introduce of infection. The operating surgeon must be aware of intracranial neural or vascular structures adjacent to the site of fistula and if there is any risk of injury this may influence the type of repair used. For example, for a lateral sphenoid defect adjacent to the optic nerve or internal carotid artery, it is safer to use a graft on the nasal surface only rather than attempt to introduce a “bath plug”. CONCLUSION The endoscopic transnasal technique for repair of anterior skull base CSF fistula is safe and effective and is now the procedure of choice for patients presenting with CSF rhinorrhoea due to anterior skull base defects. This series confirms that the “bath plug” technique of endoscopic repair is effective and reliable and may be used for a variety of types of defects. REFERENCES 1. Marshall AJ, Jones NS et al. CSF rhinorrhoea: the place of endoscopic sinus surgery. British Journal of Neurosurgery 2001; 15(1): 8–12. 2. Hughes RGM, Jones NS et al. The endoscopic treatment of cerebrospinal fluid rhinorrhoea: the Nottingham experience. The Journal of Laryngology and Otology 1997; 111(February): 125–128. 3. Burns JA, Dodson EE, Gross CW. Transnasal endoscopic repair of cranionasal fistulae: a refined technique with long-term follow-up. Laryngoscope 1996; 106: 1080–1083. 4. Hegazy H, Carrau R et al. Transnasal endoscopic repair of cerebrospinal fluid rhinorrhea: a meta-analysis. Laryngoscope 2000; 110(7): 1166–1172. 5. Badia L, Loughran S et al. Primary spontaneous cerebrospinal fluid rhinorrhoea and obesity. American Journal of Rhinology 2001; 15: 117–119. 6. Wormald PJ, McDonogh M. ‘Bath-plug’ technique for the endoscopic management of cerebrospinal fluid leaks. The Journal of Laryngology and Otology 1997; 111(November): 1042–1046. 7. Zweig JL, Carrau RL et al. Endoscopic repair of cerebrospinal fluid leaks to the sinonasal tract: predictors of success. Otolaryngology – Head and Neck Surgery 2000; 123(3): 195–201. 8. Lanza DC, O’Brien DA, Kennedy DW. Endoscopic repair of cerebrospinal fluid fistulae and encephaloceles. Laryngoscope 1996; 106: 1119–1125. 9. Chin G, Rice DH. Transnasal endoscopic closure of cerebrospinal fluid leaks. Laryngoscope 2003; 113: 136–138. 10. Aarabi B, Leibrock LG. Neurosurgical approaches to cerebrospinal fluid rhinorrhea. Ear, Nose and Throat Journal 1992; 71(7): 300–305. 11. Moseley JI, Carton CA et al. Spectrum of complications in the use of intrathecal fluorescein. Journal of Neurosurgery 1978; 48(May): 765–767. 12. Mattox DE, Kennedy DW. Endoscopic management of cerebrospinal fluid leaks and cephaloceles. Laryngoscope 1990; 100: 857–862. 13. Casiano RR, Jassir D. Endoscopic cerebrospinal fluid rhinorrhea repair: is a lumbar drain necessary? Otolaryngology – Head and Neck Surgery 1999; 121: 745–750.
Journal of Clinical Neuroscience (2004) 11(6), 597–599