Abstract
We reviewed the results of acute management of patients who had sustained a dural tear during an operation on the lumbar spine, and we attempted to determine the long-term sequelae of this complication. In the five years from July 1989 to July 1994, 641 consecutive patients had a decompression of the lumbar spine, performed by the senior one of us; of these patients, eighty-eight (14 percent) sustained a dural tear, which was repaired during the operation. The duration of follow-up ranged from two to eight years (average, 4.3 years). Postoperative management consisted of closed suction wound drainage for an average of 2.1 days and bed rest for an average of 2.9 days. Of the eighty-eight procedures that resulted in a dural tear, forty-five were revisions; these revisions were performed after an average of 2.2 previous operations on the lumbar spine, all of which resulted in a scar adherent to the dura. Only eight patients had headaches related to the spinal procedure and photophobia in the postoperative period; these symptoms resolved in all but two patients, both of whom had had a revision operation. Each of the two patients had symptoms of a persistent leak of spinal fluid and needed a reoperation for repair. Overall, seventy-six patients had a good or excellent result and twelve had a poor or satisfactory result with some residual back pain. One patient had arachnoiditis, and another had symptoms of viral meningitis one month postoperatively.A dural tear that occurs during an operation on the lumbar spine can be treated successfully with primary repair followed by bed rest. Such a tear does not appear to have any long-term deleterious effects or to increase the risk of postoperative infection, neural damage, or arachnoiditis. Closed suction wound drainage does not seem to aggravate the leak and can be used safely in the presence of a dural repair.
A dural tear is one of the most common complications of operations on the spine, with a reported prevalence of less than 1 to 17 percent in series ranging from five to 450 patients3,4,7,8. Despite the high prevalence of this complication, we are aware of no large series in the literature in which the results of treatment have been reported. The few articles on the subject all are based on fewer than twenty dural tears3,7-10.
Jones et al. found that the outcomes of seventeen operations in which a dural tear had occurred and subsequently had been treated successfully were no different than those of operations in which this complication had not occurred7. In a study by Kitchel et al., nineteen dural tears from which cerebrospinal fluid leaked postoperatively were successfully treated with closed lumbar subarachnoid drainage; however, four tears necessitated a reoperation8. On the basis of their experience with five dural tears, Eismont et al. recommended that the use of epidural and subcutaneous wound drains be avoided in order to prevent the formation of myelocutaneous fistulae3.
The current knowledge of the sequelae and treatment of dural tears is based on a few studies comprising small numbers of patients. Larger numbers of patients are needed in order to draw meaningful conclusions regarding the clinical course and final outcome when a dural tear occurs during an operation on the lumbar spine.
The purpose of the current study was to review the results of treatment of a large series of dural tears, sustained during operations on the lumbar spine that had been performed by the senior one of us (H. H. B.). Specifically, we wanted to determine the need for closed lumbar subarachnoid drainage, the prevalence of arachnoiditis, whether the duration of bed rest (average, 2.9 days) after the repair was adequate, whether the use of subfascial closed suction wound drainage was safe, and the long-term results.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopaedic Surgery, University of California at Los Angeles School of Medicine, Box 956902, Los Angeles, California 90095-6902.
‡University Hospitals Spine Institute, Department of Orthopaedic Surgery, University Hospitals of Cleveland and Case Western Reserve University School of Medicine, 11100 Euclid Avenue, Cleveland, Ohio 44106.
§Department of Orthopaedic Surgery, Washington University Medical Center, One Barnes Hospital Plaza, Suite 11300, West Pavilion, St. Louis, Missouri 63110.
The records and operative reports of 641 consecutive patients who had had a decompression of the lumbar spine between July 1989 and July 1994 were reviewed by an independent observer (J. C. W.) (Table I). All of the operations were performed by the senior one of us, who was assisted by residents and spine fellows. Of the 641 patients, eighty-eight (14 percent) sustained an intraoperative dural tear. There were thirty-four men and fifty-four women. The average age of the patients at the time of the operation was 63.6 years (range, eighteen to eighty-three years). The prevalence of dural tears varied according to diagnosis (Table I), and tears were more common in association with complex revision procedures. The duration of follow-up averaged 4.3 years (range, two to eight years). All eighty-eight patients had a lumbar decompression with a laminectomy and a foraminotomy. Fifty-four patients also had an arthrodesis, and twenty-eight of them had pedicle-screw instrumentation. Forty-five patients had a revision operation as the index procedure, after an average of 2.2 previous operations.
All but one tear was posterior or posterolateral; the exceptional tear was a ventral tear that was not repaired because of its location. All eighty-seven repairs were primary and were performed with 4-0 or 5-0 silk interlocking suture, placement of Gelfoam over the repair site, and use of a subfascial closed suction wound drain and a layered closure3. One patient who had a revision had an extensive tear, and a myofascial patch was needed to achieve a watertight closure.
Postoperatively, all eighty-eight patients were managed with the same protocol, which consisted of a short period of bed rest (average, 2.9 days) followed by walking with the aid of a physical therapist. Any symptoms of a cerebrospinal-fluid leak (positional headaches, photophobia, nausea and vomiting, clear drainage from the wound, or a fluctuant subcutaneous wound) were documented and carefully followed until resolution. Patients in whom these symptoms persisted had myelography with enhanced computed tomography to document and localize the fistula.
The use of the subfascial drain was continued for approximately two days or until the drainage was negligible. No patient was managed with a subarachnoid drain.
Detailed reviews of the charts for all of the patients were conducted to determine the method of treatment, whether there had been a persistent leak of spinal fluid, and whether a reoperation had been necessary. Residual problems were documented, and the outcome was rated. A rating of excellent indicated complete resolution of the preoperative symptoms, with no residual pain in the back or the lower extremities; a rating of good, nearly complete resolution of the preoperative symptoms, with infrequent, minor pain in the back or the lower extremities that did not necessitate narcotic medications; a rating of satisfactory, persistent, mild pain in the back or the lower extremities that was notably less severe than the preoperative level and did not necessitate narcotic medications; and a rating of poor, symptoms that were only slightly less severe or were worse than they had been preoperatively.
Eighty-six of the eighty-eight patients who had a dural tear were managed successfully with an intraoperative primary repair, and an additional patient needed a myofascial patch. The remaining patient had a tear that was not repairable because of its ventral location; however, this did not result in a persistent leak of cerebrospinal fluid in the postoperative period. A subfascial closed suction drain was used for all patients and remained in place for an average of 2.1 days (range, one to three days). No patient had a myelocutaneous fistula postoperatively. Bed rest was prescribed for all patients, and the average duration was 2.9 days (range, one to six days). The estimated operative blood loss averaged 800 milliliters. The subfascial drains averaged an output of 508 milliliters per day and a total drainage of 1062 milliliters (range, 400 to 2025 milliliters).
Of the eighty-eight patients, eight had headaches related to the operation on the spine and photophobia postoperatively, but six had resolution of these symptoms after increased intravenous administration of fluids and three days of bed rest. The two patients who had persistent symptoms of a dural leak despite three days of strict bed rest both had had a revision procedure; one had had seven operations before the revision, and the other had had five. Both had had spinal instrumentation without cross-links as part of the index procedure, whereas the six patients in whom the symptoms had resolved after bed rest had not had spinal instrumentation as part of that procedure. Both of the patients needed a reoperation for closure of the dural leak. After the exploration and repeat repair, these two patients were managed with the same protocol as had been used previously; both had a successful outcome.
Seventy-six patients had a good or excellent result, nine had a satisfactory result, and three had a poor result. No patient had a neurological deficit or an increase in the symptoms of the lower extremities as a result of the operation.
One patient had arachnoiditis before the index operation and at the latest follow-up evaluation, as confirmed by a magnetic resonance imaging study. This patient had had three operations on the lumbar spine before the index procedure. He also had a history of an injection of chymopapain and had had multiple invasive studies, including myelograms, several computed tomographic scans, and epidural injections. He did not have a persistent dural leak postoperatively and did not need a reoperation for treatment of the dural tear; the outcome was satisfactory.
One patient who had been managed successfully with a primary dural repair without the need for a reoperation had symptoms of atypical meningitis one month after the durotomy. This patient did not have a leak of cerebrospinal fluid postoperatively and was ultimately diagnosed as having viral meningitis, which resolved without additional problems.
Of the fifty-four patients who had an arthrodesis as part of the index procedure, five had a pseudarthrosis. Three of the eighty-eight patients in the series had a subsequent operation on the lumbar spine that was unrelated to the dural tear: one had a revision decompression and two had repair of a pseudarthrosis. All three patients had an excellent result.
A dural tear is one of the most common complications encountered in operations on the spine. Although the treatment of this potentially complex problem is described in the literature, these descriptions are based on relatively few studies with small numbers of patients3,7-10. Recommendations for the treatment of dural tears have included primary repair; closed subarachnoid drainage; laser tissue-welding; grafts consisting of muscle, fat, or fascia; blood patches; fibrin-adhesive or cyanoacrylate polymer sealant; application of Gelfoam to the tear; bed rest; and avoidance of the use of wound drains1-3,5,7-11. Because of the small numbers of patients in previous reports, it is difficult to compare the efficacy of the different forms of treatment; however, most authors have advocated the use of a combination of these measures.
The treatment of a dural tear without a primary repair can lead to the formation of meningeal pseudocysts, nerve-root entrapment with resultant neurological damage, a persistent leak of cerebrospinal fluid, meningitis, or arachnoiditis. There are several reports of these complications in the literature1-3,6,8,9; however, the prevalence of these problems is unknown.
To our knowledge, the largest previous study in the literature regarding the results of treatment of dural tears was that reported by Jones et al. in 19897. With use of a detailed follow-up questionnaire and an examination, those authors compared the results for seventeen patients who had sustained a dural tear during a procedure on the lumbar spine with those for age-matched controls who had not sustained a tear. They concluded that the presence of a dural tear during a spinal operation did not increase the perioperative morbidity or compromise the final result.
Goodkin and Laska, in a recent review of malpractice lawsuits involving spinal operations, found that unintended incidental durotomy was the second most common problem, accounting for twenty-three (16 percent) of 146 suits6. Those authors concluded that, because of the high frequency of legal complaints regarding dural tears, these events cannot be considered benign and may be associated with increased perioperative morbidity and long-term sequelae. Apparently, the occurrence of a dural tear generates controversy regarding the final clinical outcome of patients who have had an operation on the lumbar spine.
Eismont et al. reported on five patients who had a persistent leak of cerebrospinal fluid in association with a dural tear that had been sustained during an operation on the lumbar spine3. Those authors recommended meticulous repair of the sites of all durotomies to prevent meningitis and the formation of pseudocysts, and they advocated avoidance of the use of subfascial drains in the postoperative period to prevent the formation of cutaneous fistulae. Bed rest alone was not found to be effective.
The current study represents the largest reported series of dural tears secondary to an operation on the lumbar spine of which we are aware. A watertight primary repair with use of dural silk suture followed by an average of approximately three days of bed rest resulted in resolution of the leak of cerebrospinal fluid in eighty-six of our eighty-eight patients. The two patients who had failure after the initial primary closure and a period of bed rest were subsequently managed successfully with a repeat repair with use of the same method. Although some patients had more than three days of bed rest, no patient was managed with more than three days of bed rest for treatment of the symptoms of the dural tear alone. Longer periods of bed rest are not usually effective, and direct operative repair is recommended if leakage persists. Although no subarachnoid drains were used in the current study, their use may be appropriate in patients who have a dural tear8.
A persistent leak is suspected on the basis of an unrelenting headache, secondary to a spinal operation, that necessitates large doses of narcotic medications; severe photophobia; nausea and vomiting; increased symptoms associated with positional changes; or fluid drainage through the drain or the wound itself. The leak is confirmed with myelography and computed tomography in order to identify and localize the fistula. We prefer to use these two modalities rather than magnetic resonance imaging for the diagnosis of suspected dural leaks in the postoperative period.
The risk of meningitis or arachnoiditis appears to be very low. Meningitis usually develops in association with persistent dural leaks and cutaneous fistulae. One patient in our study had viral meningitis one month after the operation, but symptoms of a persistent dural leak did not develop postoperatively. This patient had an excellent result without any apparent residual effects from the dural tear.
Only one patient had arachnoiditis at the time of the latest follow-up, but this condition had been present before the index procedure. The dural tear in this patient was repaired primarily and did not leak in the postoperative period. The arachnoiditis did not appear to be related to the dural tear.
Despite admonitions in the literature against the use of subfascial drains3,6,7, all of our patients were managed with a closed suction drain and none had a myelocutaneous fistula. The use of a subfascial wound drain to attempt to prevent the formation of a hematoma did not result in the development of a meningeal cyst or a meningeal infection, and only two patients had a persistent dural leak. If a subfascial closed suction wound drain is believed to be necessary, its use appears to be safe even in the presence of a dural repair.
The routine use of subarachnoid drains is not recommended as most dural tears do not lead to persistent leaks. However, we do not dispute the value of these drains when a persistent dural leak cannot be repaired operatively as the drains have been shown, in previous studies, to be very effective7. Kitchel et al. reported on nineteen patients who had a postoperative cerebrospinal leak that was treated with placement of a subarachnoid drain8. This resulted in resolution of the dural leaks. The use of such a drain provides a nonoperative alternative for the treatment of postoperative leaks. In the current series, the two patients who had a persistent leak were managed with a reoperation for closure of the dural tear, and the leak resolved in both of them. A subarachnoid drain may have been a reasonable alternative in these two patients.
Both of the persistent dural leaks were in patients who had had a revision. Dural tears are more common during revision operations in patients who have epidural fibrosis and scar tissue adherent to the dura. The reasons for dural tears during primary operations include eroded or thin dura, dural adhesions, and redundant dura in patients who have tight spinal stenosis. The operative technique plays a key role in the avoidance of injury to the dural tissue.
Both of our patients who had a reoperation for closure of a dural leak had had pedicle-screw instrumentation without cross-links as part of the index procedure, whereas none of the patients who had resolution of the symptoms had had spinal instrumentation as part of the index procedure. Spinal instrumentation, especially the use of cross-links, may result in more so-called dead space surrounding the dura and may prevent the paraspinal muscles from directly tamponading a dural tear. This might be a factor in the successful resolution of a dural leak; however, the present study did not demonstrate deleterious effects from the use of spinal instrumentation.
Adequate exposure of the tear and the surrounding normal dura is necessary for the proper repair of a dural tear. Exposure of the surrounding dura did not result in spinal instability in any patient in the present study, and no patient needed a supplemental, unplanned arthrodesis as a direct result of the dural tear. The use of fine vascular needle-holders and forceps decreases the amount of exposure that is needed. Good operative lighting and magnification (with loupes or a microscope) also are recommended for suturing the tear and for protecting the neural elements from being trapped in the repair. If a patch is necessary, we use fascia from the erector spinae muscles as the fascia lata area typically is not in the operative field. The dural repair usually adds twenty to thirty minutes to the overall operative time. There is no substantial increase in blood loss or postoperative drainage from patients who have a dural tear compared with those who do not have a tear.
The long-term outcomes for our eighty-eight patients, approximately half of whom had a revision operation, are comparable with those that have been reported in the literature4,6. Seventy-six patients had a good or excellent result, nine had a satisfactory result, and three had a poor result. No patient had a neurological deficit as a result of the operation, and our patients did not have increased symptoms in the lower extremities, radiculopathy, or pain compared with patients without dural tears.
As was observed in previous investigations4,7,8, we found that an unintended incidental durotomy resulted in no substantial difference in the final outcomes of operative procedures on the lumbar spine. Therefore, we concluded that a dural tear does not adversely influence the long-term results of operations on the lumbar spine. Dural tears are more common during revisions, but they can almost always be repaired primarily, with a good or excellent outcome and without additional complications.
Cain, J. E., Jr.; Dryer, R. F.; and Barton, B. R.: Evaluation of dural closure techniques. Suture methods, fibrin adhesive sealant, and cyanoacrylate polymer. Spine,13: 720-725, 1988.13720
1988
[PubMed]
Cammisa, F. P., Jr.; Eismont, F. J.; and Green, B. A.: Dural laceration occurring with burst fractures and associated laminar fractures. J. Bone and Joint Surg.,71-A: 1044-1052, Aug. 1989.71-A1044
1989
Eismont, F. J.; Wiesel, S. W.; and Rothman, R. H.: Treatment of dural tears associated with spinal surgery. J. Bone and Joint Surg.,63-A: 1132-1136, Sept. 1981.63-A1132
1981
Finnegan, W. J.; Fenlin, J. M.; Marvel, J. P.; Nardini, R. J.; and Rothman, R. H.: Results of surgical intervention in the symptomatic multiply-operated back patient. Analysis of sixty-seven cases followed for three to seven years. J. Bone and Joint Surg.,61-A: 1077-1082, Oct. 1979.61-A1077
1979
Foyt, D.; Johnson, J. P.; Kirsch, A. J.; Bruce, J. N.; and Wazen, J. J.: Dural closure with laser tissue welding. Otolaryngol.-Head and Neck Surg.,115: 513-518, 1996.115513
1996
Goodkin, R., and Laska, L. L.: Unintended "incidental" durotomy during surgery of the lumbar spine: medicolegal implications. Surg. Neurol.,43: 4-12, 1995.434
1995
[PubMed]
Jones, A. A.; Stambough, J. L.; Balderston, R. A.; Rothman, R. H.; and Booth, R. E., Jr.: Long-term results of lumbar spine surgery complicated by unintended incidental durotomy. Spine,14: 443-446, 1989.14443
1989
[PubMed]
Kitchel, S. H.; Eismont, F. J.; and Green, B. A.: Closed subarachnoid drainage for management of cerebrospinal fluid leakage after an operation on the spine. J. Bone and Joint Surg.,71-A: 984-987, Aug. 1989.71-A984
1989
Miller, P. R., and Elder, F. W., Jr.: Meningeal pseudocysts (meningocele spurius) following laminectomy. Report of ten cases. J. Bone and Joint Surg.,50-A: 268-276, March 1968.50-A268
1968
Nash, C. L., Jr.; Kaufman, B.; and Frankel, V. H.: Postsurgical meningeal pseudocysts of the lumbar spine. Clin. Orthop.,75: 167-178, 1971.75167
1971
[PubMed]
Patel, M. R.; Louie, W.; and Rachlin, J.: Postoperative cerebrospinal fluid leaks of the lumbosacral spine: management with percutaneous fibrin glue. AJNR: Am. J. Neuroradiol.,17: 495-500, 1996.17495
1996
[PubMed]