AUTHOR(S): Shapiro, Scott A., M.D.; Scully, Thomas, M.D. Section of Neurosurgery, Department of Surgery, Indiana University Medical Center, Indianapolis, Indiana Neurosurgery 30; 241-245, 1992 ABSTRACT: One hundred and seven patients who had a lumbar subarachnoid catheter (teflon or silicone) placed for closed continuous cerebrospinal fluid (CSF) drainage between 1983-1991 are presented. Overall, the drain was successful in achieving the desired goal in 101 of 107 (94%) cases. There were no deaths. Five of 107 (5%) patients developed infections including two cases (2%) of meningitis. There were three cases (3%) of overdrainage with temporary neurologic decline, but all recovered. Five of fifteen (33%) teflon catheters required replacement because of occlusion, but only 5 of 92 (5%) silicone catheters required replacement. Transient lumbar nerve root irritation was seen in 15 of 107 (14%) patients treated for a CSF fistula, and all symptoms resolved after drain removal. CSF fistula/pseudomeningocele after spine surgery was cured by CSF drainage in 36 of 39 (92%) cases; there was a 10% incidence of infection (1 wound, 2 discitis, 1 meningitis). CSF fistula after cranial surgery was cured in 22 of 25 (87%) cases; there was 1 case of (4%) infection and 1 case (4%) of overdrainage. A drain was used to augment a tenuous dural closure in 38 patients with 100% success; no infection occurred and there were 2 cases (5%) of overdrainage. Five patients were successfully treated for traumatic CSF rhinorrhea/otorrhea without complications. The silicone catheter appears superior to the teflon catheter; however, both are simple, safe, and efficacious for the treatment or prevention of CSF fistulas. KEY WORDS: Cerebral relaxation; Cerebrospinal fluid drainage; Cerebrospinal fluid fistula; Lumbar subarachnoid catheter; Pseudomeningocele; Pituitary surgery INTRODUCTION Closed continuous drainage of cerebrospinal fluid (CSF) via a lumbar subarachnoid catheter has been used in neurosurgery for a variety of reasons. The predominant reason involved the treatment of cranial and spinal CSF fistulas. There have been few reports concerning the use of closed continuous CSF drainage via a lumbar subarachnoid catheter. Furthermore, these previous reports concerned the
use of an epidural catheter not truly designed for lumbar subarachnoid drainage and not the specially designed silicone lumbar subarachnoid drainage catheter (Cordis Corporation, Miami, FL) (1,3,5,6,8, 16-18) . Additionally, no reports have dealt with the use of lumbar CSF drainage to prevent CSF fistula developing from tenuous dural closures. Therefore, we report, in detail, our experience with closed continuous lumbar subarachnoid CSF drainage, using either a teflon epidural catheter (TC) or the specifically designed silicone subarachnoid catheters (SC) in 107 patients managed between 1983 and 1991. PATIENTS AND METHODS One hundred and seven patients hospitalized in our center during the years 1983 to 1990 had a lumbar subarachnoid catheter placed for closed continuous CSF drainage to treat or prevent a CSF fistula. The patients have been divided into four groups as shown in Table 1. Early in our experience, a marked, radiopaque TC, still in use for epidural anesthesia (continuous epidural double decker anesthesia tray, Abbott Laboratories, North Chicago, IL) today, was placed in the subarachnoid space through an 18-gauge spinal needle, using routine sterile procedure, and then was hooked up to an empty transfer pack with nonvented intravenous tubing with a macrodrip chamber. The catheter was positioned so that it coursed transversely across the back and slightly up the side of the abdomen. The catheter insertion site was treated with povidone-iodine ointment (Betadine, The Purdue Frederick Co., Norwalk, CT), and the surrounding skin was painted with mastisol. The entire catheter was covered with Op-Site adhesive drape (Acme United Corp., Bridgeport, CT). Since 1985, we have almost exclusively used the Cordis lumbar drainage set. This set consists of a larger bore radiopaque SC with depth markings, a guide wire, a 14-gauge spinal needle, and an external drainage set that allows accurate quantification of CSF drained. The catheter was dressed as above. The patients were kept at absolute bedrest but allowed to turn from side to side and could sit up to about a 45° angle in bed. The drainage chamber was adjusted with respect to the patients head so that the drainage rate could be altered and overdrainage could be prevented. The desired drainage rate was approximately 5 to 15 ml/h or 120 to 360 ml/d. Active leaks sometimes required the higher drainage rate. We had to occasionally modify the procedure for morbidly obese patients because normal length needles would not reach the lumbar subarachnoid space. In these patients, a splenic biopsy needle was used, which would allow passage of a TC through the needle (the SC was too big to fit through the needle). The external drainage system was as discussed above. Additionally, in one massively obese pituitary patient who had CSF rhinorrhea postoperatively, we had to make a small incision in the midline lumbar region down to the fascia to allow placement of a SC through a spinal needle. Both systems were left in place from 5 to 10 days
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 February 1992, Volume 30, Number 2 241 Closed Continuous Drainage of Cerebrospinal Fluid via a Lumbar Subarachnoid Catheter for Treatment or Prevention of Cranial/Spinal Cerebrospinal Fluid Fistula Experimental and Clinical Study
RESULTS The use of the lumbar subarachnoid drain was successful in treating or preventing CSF fistula in 101 of 107 (94%) patients with little morbidity, as documented in Table 2. Interestingly, transient lumbar nerve root complaints were seen in 15 of 107 (14%) patients treated for a CSF fistula, 14 of whom were treated with an SC. Only 1 of 107 (1%) patients developed a persistent spinal headache that required a blood patch. Secondary cerebrospinal fluid fistula/pseudomeningocele after spinal surgery Thirty-nine patients were treated for a pseudomeningocele/spinal fluid leak after spinal surgery, as documented in Table 3. Leakage of CSF occurred in 16 patients with intradural abnormalities despite a primary closure or dural patch graft. Of the remaining 23 cases, 19 occurred in lumbar surgeries, including nine laminectomies for stenosis/lateral recess syndromes, nine microdiscectomies, and one burst fracture. In 6 of 19 (33%) of the lumbar operations, a dural tear was identified during the operation and repaired. In 13 of 19 (66%) lumbar cases, no tear or leak was identified during the operation. In 5 of 13 (38%) of these cases, a myelogram was performed the day before surgery. The patients with cervical extradural abnormality who developed CSF problems were 2 patients with operative fractures and 2 with transoral rheumatoid arthritis and atlantoaxial instability. Three of 39 patients (8%) underwent successful subsequent operations when lumbar drainage failed to
cure their leaks. Length of follow-up ranged from 6 months to 7 years (mean, 2.8 years). Further analysis revealed that 3 of 6 (50%) TCs required replacement because of occlusion, whereas only 4 of 33 (12%) SCs required replacement, all because they came out of the subarachnoid space. Nine patients (23%), 1 with a TC and 8 with SCs, complained of transient lumbar nerve root irritation, which resolved when the drain was removed. The length of time the drain was left in place ranged from 5 to 11 days (mean, 7 days). There were 4 patients (10%) with infections, including 1 patient with a lumbar laminectomy who developed a postoperative CSF leak and a wound infection that did not respond to spinal drainage. The patient underwent subsequent surgery. The site of the leak was a myelogram needle puncture, two laminar levels above the previous surgery. Six weeks later, the patient sought treatment for Staphylococcus aureus discitis. A second patient with S. aureus discitis complicating a spinal surgery with CSF leak successfully treated by lumbar drainage was seen. Also, 1 patient with bacterial meningitis and one patient with a wound infection were seen. All of the above infections were successfully treated with antibiotics, with no long-term sequelae. Secondary cerebrospinal fluid fistula after cranial surgery Twenty-five patients were treated for a CSF fistula after cranial/transsphenoidal surgery, as documented in Table 4. The primary infratentorial abnormalities involved 13 retromastoid craniectomies for acoustic neurinomas or vestibular neuronectomies that leaked. Patients with supratentorial abnormalities included four patients with tumors involving the cribriform plate and two with transsphenoidal pituitary tumors (one microadenoma and one macroadenoma), all of whom developed CSF rhinorrhea after surgery. The length of follow-up ranged from 6 months to 7 years (mean, 3.7 years). The three (12%) patients with failures who required subsequent operations included two patients with acoustic neurinomas that had open air cells around the porous acousticus, which required bone wax closure, and one patient with cribriform plate esthesioneuroblastoma that required placement of a bone graft and fascia lata graft to stop the leak. Additionally, 2 of 5 (40%) TCs placed in the lumbar subarachnoid space had to be replaced because of occlusion and/or kinking, whereas only 1 of 18 (5%) SCs had to be replaced secondary to occlusion. Headache or nausea and vomiting could not be reliably assessed in this group because they are much too common after cranial surgery. Four cases (16%) of transient lumbar nerve root irritation occurred, all with an SC. The length of time the drain was left in place ranged from 1 to 7 days (mean, 6 days). There was 1 (4%) case of bacterial meningitis in a transsphenoidal patient whose subarachnoid drain controlled the leak. The surgery patient with esthesioneuroblastoma (4%) developed tension pneumocephalus that required drain removal, needle evacuation of the air, and another operation. Intraoperative augmentation of tenuous dural
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
for CSF leakage problems. Both systems occasionally had to be adjusted or pulled back if they stopped draining. The TC was not malleable, and any acute kink in the catheter permanently occluded flow, necessitating replacement of the catheter. Also, the TC had a smaller internal diameter that was more prone to occlusion by debris, necessitating catheter replacement. Rarely was catheter replacement necessary with the SC. Often, we clamped the system for 24 to 48 hours before removing the system. A suture was placed at the drainage insertion site after removal. All patients who underwent closed continuous CSF drainage via a lumbar subarachnoid catheter were treated intravenously with a penicillin/cephalosporin antibiotic during the course of external drainage. All patients were treated for deep vein thrombosis prophylaxis, including the subcutaneous administration of heparin and the use of antiembolic stockings. When the drain was placed to facilitate delivery of suprasellar pituitary tumors through a transsphenoidal approach, only 1 to 2 ml of spinal fluid was allowed to drain at insertion. Air was injected, 5 ml at a time, until no more tumor could be removed. Fluoroscopy demonstrated suprasellar air that moved toward the normal sellar boundary after tumor removal. The drain was left closed if no spinal fluid was seen and removed at 24 hours. If spinal fluid was seen, the drain was opened to augment dural closure.
Traumatic cerebrospinal fluid rhinorrhea/otorrhea Five patients who developed CSF rhinorrhea/otorrhea secondary to cribriform plate fracture from head injury, who had no evidence of brain hernia through the fracture, were observed for 10 to 14 days with no cessation of the leak. No antibiotics were used before drain placement. All 5 patients were subsequently treated successfully with closed, continuous CSF drainage via an SC for 7 days, as documented in Table 6. Length of follow-up ranged from 6 to 30 months for the 5 patients. There were no complications for this small group. DISCUSSION Vourc'h (18) was the first to report the use of continuous spinal CSF drainage in neurosurgery. All
reports since have involved the use of a teflon epidural catheter or a polyethylene epidural catheter placed in the lumbar subarachnoid space (1,3,5,6,8,16,17). In 1977, Findler et al. (3) reported 50 patients who underwent continuous lumbar drainage using a polyethylene epidural catheter. They were successful in treating spinal fluid leaks in 41 of 49 (84%) patients. We tended to leave the drain in place for around 7 days, whereas Findler's group tended to leave the drain in for 8 to 10 days, and they alluded to difficulties with lumbar CSF drainage from blockage of the catheter or insufficient flow in at least 10% of the patients. The blockage was caused by debris or, in cases with very high protein content, was possibly caused by the high viscosity of the fluid. Additionally, the TC is subject to kinking, which renders it functionless. Our experience is the largest reported to date and the first to include the SC system. The above results demonstrated that regardless of the catheter used, the closed, continuous drainage of CSF via a lumbar subarachnoid catheter was both effective and safe in the vast majority of cases. The SC in our experience is vastly superior to the epidural catheter for lumbar subarachnoid drainage. It is much more malleable, preventing permanent kinking, and its larger bore is less subject to occlusion or insufficient drainage caused by debris or protein. This is best demonstrated by a lower rate of catheter replacement for the SC group. Our 92% success rate for 39 spine surgery patients who developed a secondary CSF fistula/pseudomeningocele is similar to the 90% success rate reported by Kitchel et al. (6) who used spinal drainage alone to treat 17 spinal surgery patients. Both Findler et al. (3) and McCallum et al. (8) reported 100% success in 9 of these patients, using the lumbar subarachnoid drain. Both Mayfield (7) and Kitchel et al. (6) alluded to CSF fistulas/pseudomeningoceles occurring despite meticulous dural repair, and they alluded to overlooking occult dural injuries during surgery. Our experience suggests that myelograms performed the day before surgery may increase the chance of an occult leak, and we have since done them on an outpatient basis for elective cases to avoid this. The lumbar subarachnoid drain left in place for 7 to 10 days appears an ideal initial treatment for these patients, resulting in a greater than 90% success rate. We are unaware of any reported cases of overdrainage complications in patients with spinal abnormalities who had a lumbar drain. The 24% incidence of temporary nerve root irritation in our experience fortunately resolved in all cases after drain removal. It does pose significant discomfort for a few patients, requiring narcotic analgesia. This complication has not been reported to date. It seems more common with the SC and may be related to the catheter's larger diameter. The 63% incidence of transient headache and nausea and vomiting is similar to the 58% incidence reported by Kitchel et al (6). Adjusting the rate of CSF drainage, intravenous hydration, and medication are all effective in treating the complaint of spinal headache. After drain
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
closure Thirty-eight patients underwent placement of a lumbar subarachnoid drain to augment a tenuous dural closure, as documented in Table 5. Included were 19 persons who underwent transsphenoidal pituitary surgeries where free-flowing CSF was seen after tumor removal. A fascia lata and fat graft was placed in addition to the drain. This group also included nine patients with traumatic CSF rhinorrhea through a comminuted fractured cribriform plate. All of these had evidence of brain hernia through the cribriform plate defect on coronal computed tomograph. All of these patients underwent bifrontal craniotomy, intradural fascia lata grafting, and placement of additional fascia lata extradurally over the defect covering the cribriform plate region. Additionally, a bone graft was placed if there was a sizeable bony defect. An additional 3 patients with cribriform plate tumor abnormality were treated as above and augmented with a drain. Finally, 6 spine surgery patients had tenuous dural repairs augmented by a lumbar drain. The 4 TCs and 34 SCs used in this group required no replacements. Again, headache and nausea and vomiting could not be reliably attributed to the drain for this group. Two patients (5%), both with an SC, had transient lumbar nerve root irritation. The length of time the drain was left in place ranged from 2 to 6 days (mean, 5 days). Length of follow-up ranged from 6 months to 6 years (mean, 2.9 years). There were two cases (5%) of overdrainage. One of the traumatic CSF rhinorrhea patients who underwent surgery suffered significant overdrainage of essentially 400 ml within a couple of hours as a result of an error in drainage chamber placement. This patient became comatose and developed a third cranial nerve palsy. The drain was clamped, the patient intubated, and an emergent computed tomographic scan demonstrated the absence of all cisterns and no pneumocephalus. An intracranial pressure monitor was placed, and the pressure was normal. Thirty ml of sterile saline was injected thru the drain. Fortunately, the patient made a complete recovery. Additionally another CSF rhinorrhea patient who underwent surgery became lethargic secondary to a tension pneumocephalus that responded to drain clamping and needle aspiration of the air through a burr hole.
needle aspiration. Also, the 1 patient with a third nerve palsy had 30 ml of saline solution instilled via her lumbar drain before removal in an attempt to unplug the uncus. The patient did improve after saline instillation. Both of our complications occurred because of an error in drainage chamber placement, which could have been avoided. Swanson et al. (16) reported using a flow regulation of 15 ml/h of continuous spinal drainage that should eliminate the very rapid egress of spinal fluid, which can inadvertently occur with the conventional technique. Normal adult CSF production is roughly 20 ml/h; thus, even flow regulation of 15 ml/h may not completely eliminate a pressure gradient that could lead to pneumocephalus. Interestingly, all pneumocephalus complications involved communication with an air sinus. Extreme caution must be practiced for this group, and we now employ a rate of drainage of no more than 5 to 10 ml/h. Post-traumatic CSF otorrhea/rhinorrhea secondary to basilar skull fracture can be successfully treated by lumbar drainage alone. Findler et al. (3) reported in 1977 14 patients with post-traumatic CSF rhinorrhea treated initially with lumbar drainage alone for an average of 8 days. In those patients who experienced leaking within 48 hours, the drain was successful in 8 of 9 cases. In 5 patients whose leak developed after 48 hours, the drain was not successful. Additionally, Swanson et al. (16) reported two patients successfully treated with drainage alone, however, one case was complicated by pneumocephalus that had no effect on outcome. We successfully treated 5 patients who had immediate traumatic rhinorrhea/otorrhea with lumbar drainage alone for 7 days with no infections or complications. Thus, there have been no reported infections for 21 patients with traumatic CSF rhinorrhea/otorrhea treated with a lumbar subarachnoid drain. Most post-traumatic CSF leaks are immediate, and at least 85% of cases of immediate post-traumatic rhinorrhea and almost all cases of post-traumatic otorrhea will stop on their own within 1 week ((2,10, 13-15) . The reported incidence of meningitis in patients with post-traumatic CSF fistulas varies from 3 to 50%; the infection rate tends to be higher if the CSF leak persists longer than 7 days (2,10,13). Therefore, we only instituted drainage after a 7 to 10 day period of conservative management. Prophylactic antibiotics were given during the period of drainage not only in our experience but also in Findler et al.'s (3) and Swanson et al.'s (17) experience. We did not attempt to just employ drainage with any patient with radiographic evidence of brain herniation, although it is possible some of these may have healed as well and avoided a cranial procedure. A recent report recommended extracranial repair over intracranial repair for persistent traumatic CSF rhinorrhea that, in our opinion, must be individualized for each case, and this report makes no effort to resolve this issue (9) . Finally, Ommaya (11) stated that CSF drainage should not be used to treat traumatic CSF fistulas because of the high risk of reversing the flow gradient and inducing infection. The above reported experience to date suggests the risk of infection may
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
removal, spinal headache has not been a problem. Only one patient to date has required a blood patch. The real risk for these spinal surgery patients appears to be infection. Only two cases (3%) of culture positive meningitis, including one in Kitchel's report and one in our experience, of at least 59 lumbar-drained spinal surgery patients have been reported (6). Both were successfully treated with antibiotics with no sequelae. Additionally, two patients (3%) in our report had Staphylococcus aureus discitis; both were successfully treated with prolonged antibiotics and bedrest. Fortunately, these two patients made a complete recovery with no residual back pain. One patient with a wound infection was easily treated. Thus, in the reported literature, the overall risk of infection to date appears to be around 10%. What role the drain played versus the CSF leak alone in contributing to the risk of meningitis is difficult to ascertain. The reported risk of discitis in uncomplicated disc surgery has ranged from 0.27% to 3%. Thus, it appears that the use of a drain has not dramatically added to the risk of discitis (4,12) . Our 88% success rate in treating 25 patients with secondary CSF fistulas after cranial surgery using the lumbar drain is similar to the 85% success rate reported by Findler et al. in 20 patients (3). Furthermore, 31 patients who underwent a cranial procedure and had a drain placed to augment a tenuous dural closure had no CSF leaks. The above two craniotomy groups consisted of primarily transsphenoidal surgeries for macroadenomas, posterior fossa surgeries, and cases involving cribriform plate abnormality (trauma or tumor). The two main risks for the lumbar-drained cranial groups appear to be meningitis and overdrainage. We encountered only one case of meningitis (2%) in the 52 cases mentioned above. Findler et al. (3) reported one case of meningitis that preceded drain placement in 28 cases. Kitchel et al. (6) recommended the daily monitoring of CSF cell counts and laboratory values to facilitate early recognition of infection. We prefer to not violate the closed system. Prophylactic antibiotics may or may not have an effect but are routinely used by both our center and others. We encountered 3 of 55 (5%) complications resulting from CSF overdrainage, all of which were successfully treated. Findler et al's (3) series is the only other report of extensive experience with craniotomy patients who underwent lumbar drainage for an average of 10 days, and they reported no such complication in at least 28 patients. Graf et al. (5) first alerted the neurosurgical community to the problem of CSF drainage and pneumocephalus using lumbar drainage. All three reported cases in Graf et al.'s experience involved orbitofrontal abnormality with communication between the intradural compartment and the nasal air sinuses. Overdrainage occurred in all three patients, and they developed pneumocephalus, which responded to discontinuation of the drain. We also encountered two patients with orbitofrontal abnormality in whom overdrainage occurred; one had tension pneumocephalus. In addition to drain clamping, we evacuated air, if any, via percutaneous
not be as high as predicted. In conclusion, the silicone subarachnoid catheter or teflon epidural catheter employed in a lumbar subarachnoid drainage system is very simple to use and efficacious in treating or preventing CSF fistulas. With a well-informed medical and nursing staff, complications should be minimal. Received for publication, May 6, 1991; accepted, final form, August 20, 1991. Reprint requests: Scott Shapiro, M.D., Rm. 323, Wishard Memorial Hospital, 1001 W. 10th St. Indianapolis, IN 46202.
1.
2. 3.
4. 5.
6.
7. 8.
9.
10. 11.
12. 13. 14.
Aitken R, Drake C: Continuous spinal drainage in the treatment of postoperative cerebrospinal-fluid fistulae. J Neurosurg 21:275 277, 1964. Brawley B, Kelly W: Treatment of skull fractures with and without cerebrospinal fluid fistula. J Neurosurg 26:57-61, 1967. Findler G, Sahar A, Beller A: Continuous lumbar drainage of cerebrospinal fluid in neurosurgical patients. Surg Neurol 8:455457, 1977. Ford L, Key J: Postoperative infection of intervertebral disc space. South Med J 48:1295-1303, 1955. Graf C, Gross C, Beck D: Complications of spinal drainage in the management of cerebrospinal fluid fistula. J Neurosurg 54:392-395, 1981. Kitchel S, Eismont F, Green B: Closed subarachnoid drainage for management of cerebrospinal fluid leakage after an operation on the spine. J Bone Joint Surg [Am] 71:984987, 1989. Mayfield F: Complications of laminectomy. Clin Neurosurg 23:435-439, 1976. McCallum J, Maroon J, Jannetta P: Treatment of postoperative cerebrospinal fluid fistulas by subarachnoid drainage. J Neurosurg 42:434-437, 1975. McCormack B, Cooper P, Persky M, Rothstein S: Extracranial repair of cerebrospinal fluid fistulas: Technique and results in 37 patients. Neurosurgery 27:412417, 1990. Mincy J: Post-traumatic cerebrospinal fluid fistula of the frontal sinus. J Trauma 6:618622, 1966. Ommaya A: Cerebrospinal fluid fistula, in Wilkins R, Rengachary S (eds): Neurosurgery. New York, McGraw-Hill 1985, ed 1, pp 1637-1645. Pilgaard S: Discitis following removal of lumbar intervertebral disc. J Bone Joint Surg [Am] 51:713-716, 1969. Raskind R, Doria A: Cerebrospinal fluid rhinorrhea and otorrhea of traumatic origin. Int Surg 46:223-226, 1966. Rasmussen P: Traumatisk Liquorrhea: en
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
REFERENCES: (1-18)
Table 2. Overall Results for the Use of Lumbar Subarachnoid Drain to Treat or Prevent Cerebrospinal Fluid Fistula/Pseudomeningocele
Table 3. Results for the Use of Lumbar Subarachnoid Drain to Treat Cerebrospinal Fluid Fistula/Pseudomeningocele after Spinal Surgery
Table 4. Results for the Use of Lumbar Subarachnoid Drain to Treat Cerebrospinal Fluid Fistula after Cranial Surgery
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 1. Patients Treated with Closed Continuous Drainage of Cerebrospinal Fluid via a Lumbar Subarachnoid Drain
Table 5. Results for the Use of Lumbar Subarachnoid Drain to Augment Tenuous Dural Closure
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 6. Results for the Use of Lumbar Subarachnoid Drain to Treat Traumatic Cerebrospinal Fluid Rhinorrhea/Otorrhea
use of an epidural catheter not truly designed for lumbar subarachnoid drainage and not the specially designed silicone lumbar subarachnoid drainage catheter (Cordis Corporation, Miami, FL) (1,3,5,6,8, 16-18) . Additionally, no reports have dealt with the use of lumbar CSF drainage to prevent CSF fistula developing from tenuous dural closures. Therefore, we report, in detail, our experience with closed continuous lumbar subarachnoid CSF drainage, using either a teflon epidural catheter (TC) or the specifically designed silicone subarachnoid catheters (SC) in 107 patients managed between 1983 and 1991. PATIENTS AND METHODS One hundred and seven patients hospitalized in our center during the years 1983 to 1990 had a lumbar subarachnoid catheter placed for closed continuous CSF drainage to treat or prevent a CSF fistula. The patients have been divided into four groups as shown in Table 1. Early in our experience, a marked, radiopaque TC, still in use for epidural anesthesia (continuous epidural double decker anesthesia tray, Abbott Laboratories, North Chicago, IL) today, was placed in the subarachnoid space through an 18-gauge spinal needle, using routine sterile procedure, and then was hooked up to an empty transfer pack with nonvented intravenous tubing with a macrodrip chamber. The catheter was positioned so that it coursed transversely across the back and slightly up the side of the abdomen. The catheter insertion site was treated with povidone-iodine ointment (Betadine, The Purdue Frederick Co., Norwalk, CT), and the surrounding skin was painted with mastisol. The entire catheter was covered with Op-Site adhesive drape (Acme United Corp., Bridgeport, CT). Since 1985, we have almost exclusively used the Cordis lumbar drainage set. This set consists of a larger bore radiopaque SC with depth markings, a guide wire, a 14-gauge spinal needle, and an external drainage set that allows accurate quantification of CSF drained. The catheter was dressed as above. The patients were kept at absolute bedrest but allowed to turn from side to side and could sit up to about a 45° angle in bed. The drainage chamber was adjusted with respect to the patients head so that the drainage rate could be altered and overdrainage could be prevented. The desired drainage rate was approximately 5 to 15 ml/h or 120 to 360 ml/d. Active leaks sometimes required the higher drainage rate. We had to occasionally modify the procedure for morbidly obese patients because normal length needles would not reach the lumbar subarachnoid space. In these patients, a splenic biopsy needle was used, which would allow passage of a TC through the needle (the SC was too big to fit through the needle). The external drainage system was as discussed above. Additionally, in one massively obese pituitary patient who had CSF rhinorrhea postoperatively, we had to make a small incision in the midline lumbar region down to the fascia to allow placement of a SC through a spinal needle. Both systems were left in place from 5 to 10 days
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 February 1992, Volume 30, Number 2 241 Closed Continuous Drainage of Cerebrospinal Fluid via a Lumbar Subarachnoid Catheter for Treatment or Prevention of Cranial/Spinal Cerebrospinal Fluid Fistula Experimental and Clinical Study
RESULTS The use of the lumbar subarachnoid drain was successful in treating or preventing CSF fistula in 101 of 107 (94%) patients with little morbidity, as documented in Table 2. Interestingly, transient lumbar nerve root complaints were seen in 15 of 107 (14%) patients treated for a CSF fistula, 14 of whom were treated with an SC. Only 1 of 107 (1%) patients developed a persistent spinal headache that required a blood patch. Secondary cerebrospinal fluid fistula/pseudomeningocele after spinal surgery Thirty-nine patients were treated for a pseudomeningocele/spinal fluid leak after spinal surgery, as documented in Table 3. Leakage of CSF occurred in 16 patients with intradural abnormalities despite a primary closure or dural patch graft. Of the remaining 23 cases, 19 occurred in lumbar surgeries, including nine laminectomies for stenosis/lateral recess syndromes, nine microdiscectomies, and one burst fracture. In 6 of 19 (33%) of the lumbar operations, a dural tear was identified during the operation and repaired. In 13 of 19 (66%) lumbar cases, no tear or leak was identified during the operation. In 5 of 13 (38%) of these cases, a myelogram was performed the day before surgery. The patients with cervical extradural abnormality who developed CSF problems were 2 patients with operative fractures and 2 with transoral rheumatoid arthritis and atlantoaxial instability. Three of 39 patients (8%) underwent successful subsequent operations when lumbar drainage failed to
cure their leaks. Length of follow-up ranged from 6 months to 7 years (mean, 2.8 years). Further analysis revealed that 3 of 6 (50%) TCs required replacement because of occlusion, whereas only 4 of 33 (12%) SCs required replacement, all because they came out of the subarachnoid space. Nine patients (23%), 1 with a TC and 8 with SCs, complained of transient lumbar nerve root irritation, which resolved when the drain was removed. The length of time the drain was left in place ranged from 5 to 11 days (mean, 7 days). There were 4 patients (10%) with infections, including 1 patient with a lumbar laminectomy who developed a postoperative CSF leak and a wound infection that did not respond to spinal drainage. The patient underwent subsequent surgery. The site of the leak was a myelogram needle puncture, two laminar levels above the previous surgery. Six weeks later, the patient sought treatment for Staphylococcus aureus discitis. A second patient with S. aureus discitis complicating a spinal surgery with CSF leak successfully treated by lumbar drainage was seen. Also, 1 patient with bacterial meningitis and one patient with a wound infection were seen. All of the above infections were successfully treated with antibiotics, with no long-term sequelae. Secondary cerebrospinal fluid fistula after cranial surgery Twenty-five patients were treated for a CSF fistula after cranial/transsphenoidal surgery, as documented in Table 4. The primary infratentorial abnormalities involved 13 retromastoid craniectomies for acoustic neurinomas or vestibular neuronectomies that leaked. Patients with supratentorial abnormalities included four patients with tumors involving the cribriform plate and two with transsphenoidal pituitary tumors (one microadenoma and one macroadenoma), all of whom developed CSF rhinorrhea after surgery. The length of follow-up ranged from 6 months to 7 years (mean, 3.7 years). The three (12%) patients with failures who required subsequent operations included two patients with acoustic neurinomas that had open air cells around the porous acousticus, which required bone wax closure, and one patient with cribriform plate esthesioneuroblastoma that required placement of a bone graft and fascia lata graft to stop the leak. Additionally, 2 of 5 (40%) TCs placed in the lumbar subarachnoid space had to be replaced because of occlusion and/or kinking, whereas only 1 of 18 (5%) SCs had to be replaced secondary to occlusion. Headache or nausea and vomiting could not be reliably assessed in this group because they are much too common after cranial surgery. Four cases (16%) of transient lumbar nerve root irritation occurred, all with an SC. The length of time the drain was left in place ranged from 1 to 7 days (mean, 6 days). There was 1 (4%) case of bacterial meningitis in a transsphenoidal patient whose subarachnoid drain controlled the leak. The surgery patient with esthesioneuroblastoma (4%) developed tension pneumocephalus that required drain removal, needle evacuation of the air, and another operation. Intraoperative augmentation of tenuous dural
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
for CSF leakage problems. Both systems occasionally had to be adjusted or pulled back if they stopped draining. The TC was not malleable, and any acute kink in the catheter permanently occluded flow, necessitating replacement of the catheter. Also, the TC had a smaller internal diameter that was more prone to occlusion by debris, necessitating catheter replacement. Rarely was catheter replacement necessary with the SC. Often, we clamped the system for 24 to 48 hours before removing the system. A suture was placed at the drainage insertion site after removal. All patients who underwent closed continuous CSF drainage via a lumbar subarachnoid catheter were treated intravenously with a penicillin/cephalosporin antibiotic during the course of external drainage. All patients were treated for deep vein thrombosis prophylaxis, including the subcutaneous administration of heparin and the use of antiembolic stockings. When the drain was placed to facilitate delivery of suprasellar pituitary tumors through a transsphenoidal approach, only 1 to 2 ml of spinal fluid was allowed to drain at insertion. Air was injected, 5 ml at a time, until no more tumor could be removed. Fluoroscopy demonstrated suprasellar air that moved toward the normal sellar boundary after tumor removal. The drain was left closed if no spinal fluid was seen and removed at 24 hours. If spinal fluid was seen, the drain was opened to augment dural closure.
Traumatic cerebrospinal fluid rhinorrhea/otorrhea Five patients who developed CSF rhinorrhea/otorrhea secondary to cribriform plate fracture from head injury, who had no evidence of brain hernia through the fracture, were observed for 10 to 14 days with no cessation of the leak. No antibiotics were used before drain placement. All 5 patients were subsequently treated successfully with closed, continuous CSF drainage via an SC for 7 days, as documented in Table 6. Length of follow-up ranged from 6 to 30 months for the 5 patients. There were no complications for this small group. DISCUSSION Vourc'h (18) was the first to report the use of continuous spinal CSF drainage in neurosurgery. All
reports since have involved the use of a teflon epidural catheter or a polyethylene epidural catheter placed in the lumbar subarachnoid space (1,3,5,6,8,16,17). In 1977, Findler et al. (3) reported 50 patients who underwent continuous lumbar drainage using a polyethylene epidural catheter. They were successful in treating spinal fluid leaks in 41 of 49 (84%) patients. We tended to leave the drain in place for around 7 days, whereas Findler's group tended to leave the drain in for 8 to 10 days, and they alluded to difficulties with lumbar CSF drainage from blockage of the catheter or insufficient flow in at least 10% of the patients. The blockage was caused by debris or, in cases with very high protein content, was possibly caused by the high viscosity of the fluid. Additionally, the TC is subject to kinking, which renders it functionless. Our experience is the largest reported to date and the first to include the SC system. The above results demonstrated that regardless of the catheter used, the closed, continuous drainage of CSF via a lumbar subarachnoid catheter was both effective and safe in the vast majority of cases. The SC in our experience is vastly superior to the epidural catheter for lumbar subarachnoid drainage. It is much more malleable, preventing permanent kinking, and its larger bore is less subject to occlusion or insufficient drainage caused by debris or protein. This is best demonstrated by a lower rate of catheter replacement for the SC group. Our 92% success rate for 39 spine surgery patients who developed a secondary CSF fistula/pseudomeningocele is similar to the 90% success rate reported by Kitchel et al. (6) who used spinal drainage alone to treat 17 spinal surgery patients. Both Findler et al. (3) and McCallum et al. (8) reported 100% success in 9 of these patients, using the lumbar subarachnoid drain. Both Mayfield (7) and Kitchel et al. (6) alluded to CSF fistulas/pseudomeningoceles occurring despite meticulous dural repair, and they alluded to overlooking occult dural injuries during surgery. Our experience suggests that myelograms performed the day before surgery may increase the chance of an occult leak, and we have since done them on an outpatient basis for elective cases to avoid this. The lumbar subarachnoid drain left in place for 7 to 10 days appears an ideal initial treatment for these patients, resulting in a greater than 90% success rate. We are unaware of any reported cases of overdrainage complications in patients with spinal abnormalities who had a lumbar drain. The 24% incidence of temporary nerve root irritation in our experience fortunately resolved in all cases after drain removal. It does pose significant discomfort for a few patients, requiring narcotic analgesia. This complication has not been reported to date. It seems more common with the SC and may be related to the catheter's larger diameter. The 63% incidence of transient headache and nausea and vomiting is similar to the 58% incidence reported by Kitchel et al (6). Adjusting the rate of CSF drainage, intravenous hydration, and medication are all effective in treating the complaint of spinal headache. After drain
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
closure Thirty-eight patients underwent placement of a lumbar subarachnoid drain to augment a tenuous dural closure, as documented in Table 5. Included were 19 persons who underwent transsphenoidal pituitary surgeries where free-flowing CSF was seen after tumor removal. A fascia lata and fat graft was placed in addition to the drain. This group also included nine patients with traumatic CSF rhinorrhea through a comminuted fractured cribriform plate. All of these had evidence of brain hernia through the cribriform plate defect on coronal computed tomograph. All of these patients underwent bifrontal craniotomy, intradural fascia lata grafting, and placement of additional fascia lata extradurally over the defect covering the cribriform plate region. Additionally, a bone graft was placed if there was a sizeable bony defect. An additional 3 patients with cribriform plate tumor abnormality were treated as above and augmented with a drain. Finally, 6 spine surgery patients had tenuous dural repairs augmented by a lumbar drain. The 4 TCs and 34 SCs used in this group required no replacements. Again, headache and nausea and vomiting could not be reliably attributed to the drain for this group. Two patients (5%), both with an SC, had transient lumbar nerve root irritation. The length of time the drain was left in place ranged from 2 to 6 days (mean, 5 days). Length of follow-up ranged from 6 months to 6 years (mean, 2.9 years). There were two cases (5%) of overdrainage. One of the traumatic CSF rhinorrhea patients who underwent surgery suffered significant overdrainage of essentially 400 ml within a couple of hours as a result of an error in drainage chamber placement. This patient became comatose and developed a third cranial nerve palsy. The drain was clamped, the patient intubated, and an emergent computed tomographic scan demonstrated the absence of all cisterns and no pneumocephalus. An intracranial pressure monitor was placed, and the pressure was normal. Thirty ml of sterile saline was injected thru the drain. Fortunately, the patient made a complete recovery. Additionally another CSF rhinorrhea patient who underwent surgery became lethargic secondary to a tension pneumocephalus that responded to drain clamping and needle aspiration of the air through a burr hole.
needle aspiration. Also, the 1 patient with a third nerve palsy had 30 ml of saline solution instilled via her lumbar drain before removal in an attempt to unplug the uncus. The patient did improve after saline instillation. Both of our complications occurred because of an error in drainage chamber placement, which could have been avoided. Swanson et al. (16) reported using a flow regulation of 15 ml/h of continuous spinal drainage that should eliminate the very rapid egress of spinal fluid, which can inadvertently occur with the conventional technique. Normal adult CSF production is roughly 20 ml/h; thus, even flow regulation of 15 ml/h may not completely eliminate a pressure gradient that could lead to pneumocephalus. Interestingly, all pneumocephalus complications involved communication with an air sinus. Extreme caution must be practiced for this group, and we now employ a rate of drainage of no more than 5 to 10 ml/h. Post-traumatic CSF otorrhea/rhinorrhea secondary to basilar skull fracture can be successfully treated by lumbar drainage alone. Findler et al. (3) reported in 1977 14 patients with post-traumatic CSF rhinorrhea treated initially with lumbar drainage alone for an average of 8 days. In those patients who experienced leaking within 48 hours, the drain was successful in 8 of 9 cases. In 5 patients whose leak developed after 48 hours, the drain was not successful. Additionally, Swanson et al. (16) reported two patients successfully treated with drainage alone, however, one case was complicated by pneumocephalus that had no effect on outcome. We successfully treated 5 patients who had immediate traumatic rhinorrhea/otorrhea with lumbar drainage alone for 7 days with no infections or complications. Thus, there have been no reported infections for 21 patients with traumatic CSF rhinorrhea/otorrhea treated with a lumbar subarachnoid drain. Most post-traumatic CSF leaks are immediate, and at least 85% of cases of immediate post-traumatic rhinorrhea and almost all cases of post-traumatic otorrhea will stop on their own within 1 week ((2,10, 13-15) . The reported incidence of meningitis in patients with post-traumatic CSF fistulas varies from 3 to 50%; the infection rate tends to be higher if the CSF leak persists longer than 7 days (2,10,13). Therefore, we only instituted drainage after a 7 to 10 day period of conservative management. Prophylactic antibiotics were given during the period of drainage not only in our experience but also in Findler et al.'s (3) and Swanson et al.'s (17) experience. We did not attempt to just employ drainage with any patient with radiographic evidence of brain herniation, although it is possible some of these may have healed as well and avoided a cranial procedure. A recent report recommended extracranial repair over intracranial repair for persistent traumatic CSF rhinorrhea that, in our opinion, must be individualized for each case, and this report makes no effort to resolve this issue (9) . Finally, Ommaya (11) stated that CSF drainage should not be used to treat traumatic CSF fistulas because of the high risk of reversing the flow gradient and inducing infection. The above reported experience to date suggests the risk of infection may
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
removal, spinal headache has not been a problem. Only one patient to date has required a blood patch. The real risk for these spinal surgery patients appears to be infection. Only two cases (3%) of culture positive meningitis, including one in Kitchel's report and one in our experience, of at least 59 lumbar-drained spinal surgery patients have been reported (6). Both were successfully treated with antibiotics with no sequelae. Additionally, two patients (3%) in our report had Staphylococcus aureus discitis; both were successfully treated with prolonged antibiotics and bedrest. Fortunately, these two patients made a complete recovery with no residual back pain. One patient with a wound infection was easily treated. Thus, in the reported literature, the overall risk of infection to date appears to be around 10%. What role the drain played versus the CSF leak alone in contributing to the risk of meningitis is difficult to ascertain. The reported risk of discitis in uncomplicated disc surgery has ranged from 0.27% to 3%. Thus, it appears that the use of a drain has not dramatically added to the risk of discitis (4,12) . Our 88% success rate in treating 25 patients with secondary CSF fistulas after cranial surgery using the lumbar drain is similar to the 85% success rate reported by Findler et al. in 20 patients (3). Furthermore, 31 patients who underwent a cranial procedure and had a drain placed to augment a tenuous dural closure had no CSF leaks. The above two craniotomy groups consisted of primarily transsphenoidal surgeries for macroadenomas, posterior fossa surgeries, and cases involving cribriform plate abnormality (trauma or tumor). The two main risks for the lumbar-drained cranial groups appear to be meningitis and overdrainage. We encountered only one case of meningitis (2%) in the 52 cases mentioned above. Findler et al. (3) reported one case of meningitis that preceded drain placement in 28 cases. Kitchel et al. (6) recommended the daily monitoring of CSF cell counts and laboratory values to facilitate early recognition of infection. We prefer to not violate the closed system. Prophylactic antibiotics may or may not have an effect but are routinely used by both our center and others. We encountered 3 of 55 (5%) complications resulting from CSF overdrainage, all of which were successfully treated. Findler et al's (3) series is the only other report of extensive experience with craniotomy patients who underwent lumbar drainage for an average of 10 days, and they reported no such complication in at least 28 patients. Graf et al. (5) first alerted the neurosurgical community to the problem of CSF drainage and pneumocephalus using lumbar drainage. All three reported cases in Graf et al.'s experience involved orbitofrontal abnormality with communication between the intradural compartment and the nasal air sinuses. Overdrainage occurred in all three patients, and they developed pneumocephalus, which responded to discontinuation of the drain. We also encountered two patients with orbitofrontal abnormality in whom overdrainage occurred; one had tension pneumocephalus. In addition to drain clamping, we evacuated air, if any, via percutaneous
not be as high as predicted. In conclusion, the silicone subarachnoid catheter or teflon epidural catheter employed in a lumbar subarachnoid drainage system is very simple to use and efficacious in treating or preventing CSF fistulas. With a well-informed medical and nursing staff, complications should be minimal. Received for publication, May 6, 1991; accepted, final form, August 20, 1991. Reprint requests: Scott Shapiro, M.D., Rm. 323, Wishard Memorial Hospital, 1001 W. 10th St. Indianapolis, IN 46202.
1.
2. 3.
4. 5.
6.
7. 8.
9.
10. 11.
12. 13. 14.
Aitken R, Drake C: Continuous spinal drainage in the treatment of postoperative cerebrospinal-fluid fistulae. J Neurosurg 21:275 277, 1964. Brawley B, Kelly W: Treatment of skull fractures with and without cerebrospinal fluid fistula. J Neurosurg 26:57-61, 1967. Findler G, Sahar A, Beller A: Continuous lumbar drainage of cerebrospinal fluid in neurosurgical patients. Surg Neurol 8:455457, 1977. Ford L, Key J: Postoperative infection of intervertebral disc space. South Med J 48:1295-1303, 1955. Graf C, Gross C, Beck D: Complications of spinal drainage in the management of cerebrospinal fluid fistula. J Neurosurg 54:392-395, 1981. Kitchel S, Eismont F, Green B: Closed subarachnoid drainage for management of cerebrospinal fluid leakage after an operation on the spine. J Bone Joint Surg [Am] 71:984987, 1989. Mayfield F: Complications of laminectomy. Clin Neurosurg 23:435-439, 1976. McCallum J, Maroon J, Jannetta P: Treatment of postoperative cerebrospinal fluid fistulas by subarachnoid drainage. J Neurosurg 42:434-437, 1975. McCormack B, Cooper P, Persky M, Rothstein S: Extracranial repair of cerebrospinal fluid fistulas: Technique and results in 37 patients. Neurosurgery 27:412417, 1990. Mincy J: Post-traumatic cerebrospinal fluid fistula of the frontal sinus. J Trauma 6:618622, 1966. Ommaya A: Cerebrospinal fluid fistula, in Wilkins R, Rengachary S (eds): Neurosurgery. New York, McGraw-Hill 1985, ed 1, pp 1637-1645. Pilgaard S: Discitis following removal of lumbar intervertebral disc. J Bone Joint Surg [Am] 51:713-716, 1969. Raskind R, Doria A: Cerebrospinal fluid rhinorrhea and otorrhea of traumatic origin. Int Surg 46:223-226, 1966. Rasmussen P: Traumatisk Liquorrhea: en
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
REFERENCES: (1-18)
Table 2. Overall Results for the Use of Lumbar Subarachnoid Drain to Treat or Prevent Cerebrospinal Fluid Fistula/Pseudomeningocele
Table 3. Results for the Use of Lumbar Subarachnoid Drain to Treat Cerebrospinal Fluid Fistula/Pseudomeningocele after Spinal Surgery
Table 4. Results for the Use of Lumbar Subarachnoid Drain to Treat Cerebrospinal Fluid Fistula after Cranial Surgery
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 1. Patients Treated with Closed Continuous Drainage of Cerebrospinal Fluid via a Lumbar Subarachnoid Drain
Table 5. Results for the Use of Lumbar Subarachnoid Drain to Augment Tenuous Dural Closure
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 6. Results for the Use of Lumbar Subarachnoid Drain to Treat Traumatic Cerebrospinal Fluid Rhinorrhea/Otorrhea
