Intraoperative drainage of cerebrospinal fluid (CSF) is frequently performed during neurosurgery to relax the brain, promote surgical exposure, and improve visualization of intracranial contents. Neurosurgical procedures that frequently use intraoperative CSF drainage include transsphenoidal pituitary surgery, repair of aneurysms of the circle of Willis, and orbital decompressions . In addition, neurosurgical procedures occasionally result in CSF leaks for which continuous spinal drainage is used postoperatively to promote dural healing . Neurosurgical patients requiring intraoperative lumbar spinal fluid drainage typically undergo a needle or catheter placement after general anesthesia has been induced. The needle or catheter is placed after anesthetic induction to avoid provoking undue stress or discomfort that may increase intracranial pressure. Furthermore, drain placement may be technically difficult or hazardous in patients rendered confused or combative by their neurologic disease. However, although the placement of spinal drainage catheters in anesthetized patients is well accepted, the performance of regional anesthesia on patients already receiving general anesthesia is often debated because of the theoretical increased risk of neurologic complications.
According to the ASA Closed Claims Study, injury to the lumbosacral nerves is the third most common anesthesia-related peripheral nervous system injury . Needle-related neural trauma, local anesthetic toxicity, infection, and spinal cord ischemia have been implicated in causing neural dysfunction after regional anesthesia [4,5]. In the Closed Claims Study , all cases of lumbosacral nerve injury with an identifiable anesthetic etiology were related to regional anesthesia techniques that elicited pain or paresthesias on needle insertion or local anesthetic injection. Several additional studies have demonstrated that a neurologic deficit is most likely to develop in the same dermatome as the pain or paresthesia [6,7]. Because heavily sedated or anesthetized patients are unable to report pain or paresthesias during needle or catheter placement, performance of a subarachnoid or epidural injection in this patient population is controversial .
The risk of nerve injury from intrathecal needle or catheter insertion, with or without injection of local anesthetics, is unknown. In this study, we assessed the frequency of neurologic complications caused by CSF drain placement in anesthetized patients undergoing transsphenoidal surgical procedures.
This study was approved by our institutional review board. The medical records of all patients undergoing transsphenoidal surgical procedures performed with intraoperative CSF drainage between January 1, 1985 and May 31, 1997 were identified from a computerized database of surgical cases.
General demographic information included height, weight, and dates of birth and surgery. The indication for surgery and preexisting diagnoses (if present) of acromegaly, Cushing's disease, diabetes, lumbosacral spinal cord disease, or lower extremity dysfunction were recorded. Diabetes was defined as a history of receiving insulin or an oral hypoglycemic drug. Acromegaly and Cushing's disease diagnoses were obtained from preoperative endocrinologic evaluations. Lumbosacral spinal cord disease was considered present if the patient had a history of herniated nucleus pulposus or disc, spinal stenosis, myelopathy, radiculopathy, paraparesis, paraplegia, cauda equina syndrome, neurogenic bowel or bladder, spinal cord tumor, or previous surgery to the lumbosacral spine. Lower extremity dysfunction was considered present if a history of any of the following could be elicited from the medical record: 1) pain, numbness, weakness, or a decrease in sensation to pain, touch, or temperature in the buttocks, perineum, groin, pelvic region or lower extremities; 2) neurogenic bowel or bladder; or 3) peripheral neuropathy or decrease in sensation and/or strength of the lower extremities, typically in a stocking-like distribution. Peripheral neuropathies or skeletal muscle weakness associated with endocrine disorders (e.g., acromegaly, Cushing's disease, and diabetes mellitus) were also documented preoperatively.
The technique of spinal drain insertion involved placing the patient in the lateral decubitus position after the induction of general anesthesia. The spinal drain was then inserted using an aseptic technique at the L2-3, L3-4, or L4-5 level. Access to the CSF was gained via a 19-gauge malleable needle (Hingson-Edwards continuous caudal anesthesia needle; Becton Dickson, Rutherford, NJ), a standard epidural catheter (20-gauge nylon closed-end epidural catheter; Abbott Laboratories, North Chicago, IL) placed through an 18-gauge epidural needle, or a specially designed 16-gauge CSF drainage catheter (Cordis lumbar drain; Cordis Corporation, Miami, FL) inserted with a 14-gauge epidural needle. The patient was then returned to the supine position, and drain patency was confirmed. Inadequate CSF flow was corrected by drain manipulation or replacement. With the patient in the supine position, the malleable needle protruded from the back into a cut-out area of mattress specially designed for this purpose. Although many factors influence the choice of CSF drainage technique, a reliable generalization about the CSF drainage practice at the Mayo Clinic is that malleable needles are used when CSF drainage is required exclusively during the intraoperative period and that subarachnoid catheters are placed when CSF drainage is expected to continue into the postoperative period. Each chart was examined to ensure that CSF drain insertion was performed after anesthesia induction. The type of CSF drain inserted was determined from the anesthesia record. Attempts at multiple vertebral levels and presence of blood in the CSF were noted. Use of the spinal drain for air injection, CSF drainage, or medication infusion was recorded.
Postoperative hospital notes were examined for evidence of new neurologic changes manifested by sensory or motor disturbances in the buttocks, perineum, groin, pelvic region, or lower extremities. Presence of postdural puncture headache (PDPH) and/or performance of an epidural blood patch were also recorded. Records from follow-up office visits within 1 yr of surgery were examined for new patient complaints or physical findings of lower extremity dysfunction. All cases of possible neurologic dysfunction related to spinal drainage during the postoperative hospital stay and during any subsequent follow-up visits were reviewed by a single neurologist.
The percentage of patients experiencing neurologic complications (corresponding exact 95% confidence interval [CI]) was calculated. The percentage of patients experiencing PDPH and those requiring an epidural blood patch was also calculated.
Baseline demographics and procedural characteristics from 513 patients undergoing 530 procedures are listed in Table 1. Neurologic complications are presented in Table 2.
No neurologic deficits attributable to spinal drain insertion (0%, 0.0%-0.7%) were detected in the immediate postoperative period or at postoperative visits within 1 yr of surgery. There were three patients with neurologic complications not related to CSF drainage. One patient complained of localized tenderness for 2 days at the CSF drain insertion site; however, no lower extremity symptoms were present, and the pain resolved uneventfully. One patient developed mild sensory loss of the entire left lower extremity 2 mo after an incomplete resection of a clival chordoma, attributed to tumor progression. Finally, the third patient developed meningitis 7 days after a transsphenoidal hypophysectomy that was complicated by a CSF leak. All bacterial cultures obtained from this patient, including blood and CSF, were negative, and the patient recovered with empiric antibiotic therapy. Given the history of the CSF leak (a recognized risk factor for the development of meningitis) and the lack of a spinal or epidural abscess in this patient, the cause of the infection was most likely not the CSF drain.
PDPH developed in 13 patients (2.5%, 1.3%-4.2%), 7 of whom required an epidural blood patch (1.3%, 0.5%-2.7%).
Neurological injury caused by needle insertion into the central neuroaxis is frequently heralded by pain, paresthesias, or movement. All of these warning signs may be masked by general anesthesia, which makes needle insertion in anesthetized patients controversial. This study examined the neurologic consequences from the insertion of lumbar subarachnoid drains in anesthetized patients undergoing neurosurgery.
The transsphenoidal approach is routinely used to gain access to the region of the sella turcica . By virtue of their anatomic location, lesions of the sella do not typically cause focal neurologic impairment of the lower extremities, nor does their surgical excision usually violate intracranial structures that would result in lower extremity neurologic compromise. Additionally, patients are positioned supine for transsphenoidal surgery, a position that carries little risk of lower extremity neurologic injury. Therefore, because transsphenoidal procedures frequently require spinal drainage and have little propensity for causing lower extremity neurologic dysfunction, they represent an excellent means of ascertaining the frequency of neurologic injury attributable to placing needles into the lumbar subarachnoid space of anesthetized patients.
Specific complications caused by spinal drainage seem to be infrequent but, because of their involvement with the central or peripheral nervous system, may be a source of significant morbidity or mortality. Messick et al.  found no evidence of neurologic dysfunction attributable to 19-gauge malleable needle placement in anesthetized subjects during a review of 54 transsphenoidal procedures. Fishman et al.  found no neurologic sequelae from the subarachnoid placement of epidural catheters under local anesthesia in a retrospective analysis of 23 patients who developed CSF leaks after acoustic neuroma resection. Likewise, Kitchel et al.  placed lumbar epidural catheters in 19 awake patients to treat CSF leaks without neurologic sequelae caused by the insertion. However, Roland et al.  reported radicular symptoms in 3 of 32 patients after awake subarachnoid drain insertion using an epidural catheter. Two patients had radicular pain at the time of catheter insertion that was transient in one case but that persisted until drain removal in the other. The third patient developed radicular pain and urinary retention 24 h after drain insertion. The pain abated several days after drain removal, but lower extremity weakness persisted for several months, and magnetic resonance imaging revealed an abnormality of the L5 nerve root. Matricali  reported 200 cases of lumbar subarachnoid insertion of epidural catheters to treat or prevent postoperative CSF fistulae. One patient in the series developed long-standing L5 nerve root injury after a CSF drain was inserted under general anesthesia.
Artru and Katz  compared rigid and malleable 18-gauge spinal needles with standard 20-gauge epidural catheters and 16-gauge Cordis lumbar catheters for their ability to drain CSF after insertion into anesthetized patients undergoing intracranial surgery. Based on the frequent need for repositioning and appearance of gross blood, Artru and Katz  postulated that both the malleable and rigid spinal needles produced more trauma than subarachnoid catheters; however, none of the 120 patients in their study developed neurologic sequelae aside from PDPH. On the contrary, Dripps  compared the placement of 1921 20-gauge rigid spinal needles, 1107 19-gauge malleable needles, and 506 3.5F catheters via 16-gauge Huber point needles for the administration of spinal anesthesia. The incidence of paresthesis and bloody taps was greater for subarachnoid catheters (33% and 4.4%, respectively) than for rigid (13% and 2.2%, respectively) or malleable (6% and 1.4%, respectively) spinal needles. Dripps  concluded that the catheter technique produced more local trauma but made no comment on the presence of neurologic sequelae in the postoperative period.
Injury to the spinal cord and lumbosacral nerves caused by CSF drainage has a corollary in the neurologic complications caused by spinal anesthesia because both CSF drainage and spinal anesthesia require the introduction of a needle into the subarachnoid space. All patients in our study were subjected to risks of trauma, infection, and ischemia similar to those undergoing spinal anesthesia. However, in contrast to patients receiving spinal anesthesia, patients in our study were neither awake at the time of needle insertion nor subjected to the potentially neurotoxic effects of intrathecal local anesthetics or vasoconstrictors.
Spinal anesthesia may cause neural injury via ischemia, trauma, infection, and neurotoxicity [4,5]. Dripps and Vandam  uncovered new motor weakness or sensory changes in 73 of 10,098 (0.72%) patients who had received spinal anesthesia. Horlocker et al. [4,5] retrospectively analyzed the neurologic complications resulting from single-dose and indwelling catheter techniques for the administration of spinal anesthesia. The single-dose technique was associated with neurologic complications in 6 of 4767 (0.13%) patients, whereas the catheter technique produced complications in 4 of 603 (0.66%) patients. In addition, Auroy et al.  identified 24 cases of neurologic injury in a prospective survey of 40,640 spinal anesthetics (0.06%). Likewise, Dahlgren et al.  found seven instances of neurologic dysfunction in a retrospective review of 8501 spinal anesthetics (0.08%). The 0% observed rate of neurologic injury from the current study may seem lower than the previously reported rates observed for spinal anesthesia. It is tempting to speculate that the reason for this difference in the rates of neurologic complications could be the lack of neurotoxic effects from CSF drainage. However, based on the exact 95% CI of 0.0%-0.7%, the upper limit of the true rate of neurologic injury from CSF drainage is potentially similar to the rate of neurologic injury from spinal anesthesia as determined by the other investigators [4-6,16]. This must be interpreted with caution because, despite the apparent similarities between CSF drainage and spinal anesthesia, they may produce neurologic injury by different means or currently unrecognized mechanisms.
Although it is unlikely that a prospective randomized study evaluating the risk of neurologic complications after neuraxial block in anesthetized versus awake patients will be proposed, such an investigation could be performed in patients undergoing CSF drain insertion. However, a very large sample size would be required. Using 0.1% as an estimate of the frequency of neurologic injury from awake CSF drain insertion, a total of 4000 subjects (2000 anesthetized, 2000 awake) would be required to provide 80% power to detect a fivefold increase in the frequency of neurologic injury (i.e., any rate >0.5%) for CSF drain insertion under anesthesia. Detecting an increase in neurologic injury of less than fivefold would require that even more patients be studied. These values are calculated based on a two-sample study design using a one-sided alpha = 0.05 level test. A study of this magnitude may be possible only if performed in a multi-institutional format.
Admittedly, the retrospective nature of this study lends itself to an underestimation of the true rate of occurrence of neurologic injury caused by CSF drainage or the development of PDPH. However, we believe that a significant neurologic lesion to the lower extremities would have been detected and duly noted during the postoperative hospital stay or in subsequent clinic visits. Our ability to detect neurologic dysfunction is supported by the three patients we identified who manifested neurologic injuries unrelated to CSF drainage and those patients who developed PDPH.
In summary, a retrospective analysis of 530 transsphenoidal surgical procedures in which a lumbar subarachnoid CSF drain was inserted after anesthetic induction failed to detect a single case of lower extremity neurologic dysfunction attributable to the CSF drain. Studies of the neurologic complications from spinal anesthesia have produced rates of neurologic injury similar to those found in the current study [4-7,16].
1. Michenfelder JD, Gronert GA, Rehder K. Neuroanesthesia. Anesthesiology 1969;30:65-100.
2. Graf CJ, Gross CE, Beck DW. Complications of spinal drainage in the management of cerebrospinal fluid fistula. J Neurosurg 1981;54:392-5.
3. Kroll DA, Caplan RA, Posner K, et al. Nerve injury associated with anesthesia. Anesthesiology 1990;73:202-7.
4. Horlocker TT, McGregor DG, Matsushige DK, et al. Neurologic complications of 603 consecutive spinal anesthetics using macrocatheter and microcatheter techniques. Anesth Analg 1997;84:1063-70.
5. Horlocker TT, McGregor DG, Matsushige DK, et al. A retrospective review of 4767 consecutive spinal anesthetics: central nervous system complications. Anesth Analg 1997;84:578-84.
6. Auroy Y, Narchi P, Messiah A, et al. Serious complications related to regional anesthesia: results of a prospective survey in France. Anesthesiology 1997;87:479-86.
7. Dripps RD, Vandam LD. Long-term follow-up of patients who received 10,098 spinal anesthetics. JAMA 1954;156:1486-91.
8. Bromage PR. Neurological complications of subarachnoid and epidural anaesthesia. Acta Anaesthesiol Scand 1997;41:439-44.
9. Messick JM Jr, Laws ER Jr, Abboud CF. Anesthesia for transsphenoidal surgery of the hypophyseal region. Anesth Analg 1978;57:206-15.
10. Fishman AJ, Hoffman RA, Roland JT Jr, et al. Cerebrospinal fluid drainage in the management of CSF leak following acoustic neuroma sutgery. Laryngoscope 1996;106:1002-4.
11. Kitchel SH, Eismont FJ, Green BA. Closed subarachnoid drainage for management of cerebrospinal fluid leakage after an operation on the spine. J Bone Joint Surg Am 1989;71:984-7.
12. Roland PS, Marple BF, Meyerhoff WL, Mickey B. Complications of lumbar spinal fluid drainage. Otolaryngol Head Neck Surg 1992;107:564-9.
13. Matricali B. Controlled spinal drainage in the treatment of CSF fistulae. J Neurosurg Sci 1980;24:147-9.
14. Artru AA, Katz RA. Comparison of spinal needles, epidural catheters, and cordis lumbar catheters for intraoperative removal of cerebrospinal fluid. Neurosurgery 1988;22:101-4.
15. Dripps RD. A comparison of the malleable needle and catheter technics for continuous spinal anesthesia. N Y State J Med 1950;50:1595-9.
© 1999 International Anesthesia Research Society
16. Dahlgren N, Tornebrandt K. Neurological complications after anaesthesia: a follow-up of 18,000 spinal and epidural anaesthetics performed over three years. Acta Anaesthesiol Scand 1995;39:872-80.