Continuous Intracisternal and High Cervical Intrathecal Bupivacaine Analgesia in Refractory Head and Neck Pain

Between February 5, 1990, and April 1, 1995, 13 patients, 8 men and 5 women aged 32-84 yr (mean 73 yr), with refractory pain from the head, face, mouth, neck, and upper extremities (Table 1) were treated with continuous intracisternal infusions of bupivacaine.**

Patients referred to our pain section were consecutively included in the study when fulfilling the following criteria: pain located in the head, face, mouth, neck, and shoulders (n = 13); failure of other methods to provide acceptable pain relief (n = 12); and morphine intolerance and/or unacceptable side effects (intractable constipation, hallucinations, confusion, nausea and vomiting) from opioids (n = 8). Depression, senility, and associated severe physical conditions were not considered contraindications to intracisternal bupivacaine treatment. Moribund patients and patients with overt psychoses were excluded from the study.

The materials used for the catheterization of the cerebellomedullary cistern (cisterna magna) were identical to those used for catheterization of the spinal subarachnoid space at lower levels [12]: a 9-cm 17-G Tuohy needle with a sharp bevel; a clear nylon 1.1 mm OD. 900-mm 18-G catheter (Portex 100/382/116, Hythe, Kent, England), with a closed, rounded tip and three side holes; a 0.22-micro Meter Millipore filter (Vygon 809.22, Vygon, Ecouen, France); a tunneling instrument (35-cm; Portex 04-380-999); and stainless steel multifilament threads (Stahldraht-00, Multifilament-DX-4785, Dekmatel, Hamburg, Germany) for fixation of the catheter hub. Other materials used in the procedure have been reported in detail elsewhere. [12].

Insertion of the Intracisternal Catheter. The intracisternal cannulations were performed under sterile conditions in the operating room, under monitoring of the electrocardiogram, heart rate, and oxygen saturation (pulse oximetry) and with the administration of 2 l/min oxygen by nasal catheter. Antibiotics were given prophylactically.

Suboccipital Approach. In one patient, a 16-G Tuohy needle was introduced medially between the atlas and the occiput. The cisterna magna was located at a depth of [nearly equal] 6 cm. An 18-G silicone elastomer catheter provided with a plastic stylet was advanced through the needle, [nearly equal] 10 cm from the skin and [nearly equal] 4 cm intracisternally. The catheter was tunneled subcutaneously and fixed at the tunnel exit as described below.

Low Cervical/High Thoracic Approach. The low cervical/high thoracic approach (C7-T5; Figure 1) was used in 11 patients. In 10 patients, the intracisternal catheterization was performed in the "head-up" ([nearly equal] 30 degrees) lateral recumbent position. In one patient with severe respiratory insufficiency, the intracisternal catheter was inserted with the patient in the sitting position. The spinous processes of the C7 and T1-T5 vertebrae were marked with a dermographic pencil on the patient's skin and the body of the C1 vertebra with a small, round lead shot (Figure 2), with the aid of C-arm fluoroscopy. After dural puncture, a nylon 18-G Portex catheter was threaded through the Tuohy needle. The catheter, filled with Omnipaque (Nycomed, Oslo, Norway) or with an x-ray-opaque guidewire, was advanced cranially into the subarachnoid space, until its tip reached the C1 vertebral body (Figure 1and Figure 2).

Low Thoracic Approach. A low thoracic approach (T7-T9) was used in two patients after unsuccessful attempts to insert the catheter at the C7-T7 interspaces. The Portex catheters, with a guidewire (n = 1) or with contrast in the lumen (n = 1), were advanced cranially under C-arm fluoroscopic control to the C2/C1 vertebrae (42 and 33 cm from the skin, including 38 and 28 cm intrathecally).

The catheter was tunneled paravertebrally, over the shoulder, and parasternally, with the tunnel exit at the level of the third chondrocostal junction in nine patients. Because of neoplastic infiltration of the soft tissues of the neck, the tunnel course was directed caudally to the hypochondrium in four patients.

The catheter was secured as previously described [12]: a split silicone cuff was passed around the catheter immediately below the tunnel exit and fixed to the skin by two monofilament (Novafil-00) sutures, taking a generous segment (approximately 2 cm breadth and 1 cm depth) of the integument. The sutures were knotted without causing skin ischemia, passed over the cuff, and knotted tightly to prevent to-and-fro movements of the catheter without occluding it. The catheter hub was fastened to the skin by two steel multifilament sutures (Stahldraht-00, Multifilament-DX-4785, Dekmatel, Hamburg, Germany) taking [nearly equal] 3 cm breadth and [nearly equal] 1 cm depth of the skin (Figure 3).

The skin incisions at the insertion site and along the tunnel were closed with monofilament sutures (Novafil-000) taking the whole thickness of the skin, up to the fascia. An antibacterial filter filled with 0.5% bupivacaine was connected to the catheter hub and capped. Absorbent self-adhesive compresses were applied to the skin incisions and their adhesion to the integument increased with transparent self-adhesive films. A gauze compress was applied between the Millipore filter and the skin. The tunnel exit was covered with a split absorbent compress, and the patient was turned to the supine position. After bupivacaine bolus tests in some patients (see below and Table 2), a Pharmacia Deltec patient-controlled analgesia pump (St. Paul, Minnesota) was connected to the filter. Finally, the whole tunnel exit (catheter hub, Millipore filter, and the proximal 2-4 cm of the connecting tubing) was covered with a 12.7 x 17.8-cm self-adhesive film (Bioclusive, Johnson & Johnson, New Brunswick, NJ).

Postcatheterization Care. The patients were supervised for 24 h in the postoperative ward. Heart rate, blood pressure, the electrocardiogram, pulse oximetry, respiration rate, and 24-h diuresis were recorded. Nasal oxygen 1-4 l/min (median 2 l/min) was given for 10-12 h to nine patients.

The antibiotics were continued for 3 days. The skin sutures (except those securing the catheter) were removed after 7-10 days. The dressing at the tunnel exit was changed at 48-72 h and subsequently every 7-10 days. Chlorhexidine gluconate (0.5%) in 60% alcohol was used for skin antisepsis. A fine layer of Inotyol (ichthammol 1.5% in a zinc oxide-titanium oxide-lanolin-Vaseline ointment base; Selena Lakemedel, Sundbyberg, Sweden) and an absorbent compress (Mepore, Molnlycke, Sweden) were applied over the tunnel exit and the steel suture points. The absorbent compress was kept in place by a transparent, self-adhesive dressing (Bioclusive). The transparent dressing covered the tunnel exit, the catheter hub, the filter, and the proximal 2-4 cm of the connecting tubing. The infusion system was changed by the ward or district nurse, when close to empty, within a month. The antibacterial filter was changed within a month or longer by pain nurses educated in the care of the intrathecal catheters. During filling and refilling of the infusion containers and exchange of the filter, all contact between the operator's hands and plunger shafts, nozzles of the syringes, and connections of the cassettes and needles were avoided. No other aseptic precautions (e.g., wearing of sterile rubber gloves or filling and refilling of the syringes and cassettes under a hood) were undertaken. [13].

In case of dysfunction occurring with the catheter and/or pump or inadequate pain relief, the patients, their next of kin, or the nurses on duty could contact the doctor in charge (I.C. or P.N.).

Intracisternal Treatment Schedule. Plain isotonic and isobaric (specific gravity 1.000 at 37 degrees C) bupivacaine solution 5 mg/ml (Marcaine, ASTRA, Sodertalje, Sweden) was used in all patients. The concentration was the same for all patients throughout the treatment.

Intracisternal Bupivacaine Test. In one patient, a test dose of 5 mg bupivacaine was given by direct puncture of the cisterna magna (the patient with a suboccipital approach, whose data were not included in the study). Six patients received test doses (1.5-5.0 mg) through the catheter, immediately before connecting the infusion pump (Table 2). The test bolus doses were omitted in the remaining seven patients because of their poor general condition or being too affected by their opioids and/or premedication. Another patient received an intracisternal bolus dose of 7.5 mg bupivacaine after accidental arrest of the infusion due to disconnection of the catheter from its hub and recurrence of severe pain on the evening of the insertion day.

Intracisternal Treatment. Intracisternal treatment continued with 0.5% bupivacaine infusion, usually at a rate of 0.2 ml (1 mg)/h, with optional bolus doses by patient-controlled analgesia. On-demand doses were initially less than 1.5 mg of bupivacaine, with lockout intervals of 15-60 min, and could be administered 1-4 times/h. The daily doses were gradually increased (by increasing the hourly rate, the on-demand dose, and/or the potential number of doses per hour), until adequate pain relief was obtained (VAS^mean0-1).

Associated Analgesics and Sedatives. The patients had ad libitum access to nonopioid analgesics/sedatives and to opioids administered by various routes (oral, enteral, and/or parenteral), until they obtained acceptable pain (VAS^mean0-1) and anxiolytic relief. These drugs were prescribed by the doctors in charge of the treatment (I.C. or P.N.).

Assessment of Efficacy of the Intracisternal Pain Treatment. The efficacy of the intrathecal pain treatment was evaluated by:

Self-reported Pain Intensity and Self-reported Pain Relief. Maximal, mean, [14]and minimal VAS daily pain scores were used to express pain intensity on a 10-point scale (visual analog scale; Apotekarnas Central Organisation AB/Kabi Pharmacia AB, Solna, Sweden). For the patients treated in our hospital, the VAS scores were obtained every day, at a visit once a day between 6 and 8 PM, when the patients reported the maximal and minimal intensity of the pain experienced during the last 24 h. For the patients treated in other hospitals or at home, the daily pain scores were reported daily by telephone at the same time as above to one of the authors (I.C. or P.N.) by the patients, their next of kin, or the nurses on duty.

These scores were expressed as daily scores of the day before the start of the intracisternal bupivacaine treatment, 2 days after the start (initial values), 2 days before treatment termination (final values), and as a mean of all values for the whole period (whole period, also labeled "usual pain" [14]).

Scores. The following were noted: (1) nocturnal sleep pattern scores (0-5): 0 = coma; 1 = < 2 h of uninterrupted sleep; 2 = > 2 h; 3 = > 4 h; 4 = > 6 h; and 5 = 7-8 h; and (2) scores of patient's gait pattern and ambulation (0-5): 0 = bedridden; 1 = could be moved; 2 = could move independently, e.g., with a wheelchair; 3 = could walk with crutches; 4 = could walk without help: and 5 = returned to normal activities.

Total Opioid Daily Dosage. Total opioid daily dosage, i.e., the sum of the opioid doses administered daily by all routes, expressed as mg parenteral morphine-eq per day. [15].

Effects of the Intracisternal Treatment on the Phrenic Nerve Function. Effects of the intracisternal treatment on the phrenic nerve function were indirectly estimated from the oxygen saturation (pulse oximetry) during the first 12-24 h and from any gross impairment of ventilatory function (e.g., clear respiratory distress with the use of the accessory respiratory muscle, Litten's sign, [16]presence of a scaphoid abdomen, not protruding on inspiration).

Statistics. The data were processed with a StatView statistical program (Abacus, Berkeley, CA). Descriptive and inferential statistics, including one-factor analysis of variance and a nonparametric test (Wilcoxon's signed-rank test) for repeated measures, were used. P < 0.05 was accepted as significant.

Ethics. The patients and their families were given verbal and written information about the intracisternal treatment, its possible complications, and other potential inconveniences of the therapy. Informed consent was received from all patients. Institutional approval for the study was obtained.

Intracisternal Catheterization. Suboccipital Approach. Introduction of the soft silicone catheter and withdrawal of the stylet proved to be cumbersome. The catheter became spontaneously dislodged into soft tissue 2 days after its insertion. Because of problems with the catheterization, the early dislocation, and the problems of pain evaluation due to this, we abandoned the suboccipital approach and the silicone catheter after this single case. The case was not included in the statistical data with the other patients.

High Thoracic/Cervical and Low Thoracic Approach. Insertion of the catheters appeared more complicated in patients with malignant than in those with nonmalignant pain, with more difficult dural punctures and more multiple dural punctures, accidental vascular punctures, and blood-stained cerebrospinal fluid. The differences did not reach statistical significance. There were no difficulties with the insertion of the catheters into the intrathecal space, even with long distances (33 and 42 cm) from the insertion site.

Catheter Removal. In the four observations (1, 5, 10, and 11) of catheter removal, no difficulties were encountered on withdrawal by pulling it after removal of the skin sutures securing the catheter. Neither cerebrospinal fluid leakage or subsequent headache were observed after catheter removal.

Immediate Postcatheterization Period. The results of the intracisternal bupivacaine tests with respect to doses, analgesia, and adverse effects are summarized in Table 2. The results of intracisternal bupivacaine infusion (bupivacaine doses and adverse effects) during the first 24 h are presented in Table 3.

Intracisternal Pain Treatment and Assessment of Its Efficacy. The data pertinent to intracisternal bupivacaine pain treatment are presented in Table 4. All patients had the intracisternal treatment stopped or were dead at the termination of the study. No death was attributed to the intracisternal pain treatment. The 13 patients were treated for 3-182 days (median 37, total 712 days), 3 being treated at home for 10-112 days (median = 88, total = 210 days).

In one patient, the efficacy of the treatment could not be estimated because of advanced senility and pain behavior. Assessment of treatment efficacy was difficult (but not impossible) in another four patients, including one with blindness and deafness because of chronic yeast meningitis, in whom the only means of communication was by writing letters in his palms, two patients with permanent tracheostomies speaking with a device amplifying their cervical voice, and one patient with occurrence of new pain in other parts of the body (due to new metastases) not relieved by intracisternal bupivacaine.

Self-reported Pain Intensity and Self-reported Pain Relief. Eleven of the 13 patients obtained acceptable pain relief (average pain intensity 0-2; Figure 4). In one patient with refractory pain from a malignant melanoma of the head and neck, the VAS^meancould not be reduced to an acceptable level (less or equal to 2), and high doses of intravenous morphine were added. In another patient with severe depression and occurrence of pain from new metastases in other parts of the body, the VAS^meanscores (5) remained unchanged.

Scores. The nocturnal sleep pattern scores increased significantly, corresponding to a prolongation of sleep from 2 h before to greater or equal to 6 h (medians) during the intracisternal period, whereas the gait ability pattern scores did not immediately deteriorate post-procedure (implying a procedure- or drug-related complication) but did so over a slow, disease-related time course (Figure 5).

Daily Dosages of Opioids and Intracisternal Bupivacaine. The daily doses of total opioids, expressed as parenteral morphine equivalents, and of intracisternal bupivacaine are illustrated in Figure 6. Particulars of the intracisternal treatment, such as programming of the pumps, accumulated bupivacaine doses, treatment duration, the patients' functional state (daily activities possible during treatment), and associated pharmacologic treatment are summarized in Table 4.

Total Opioid Daily Doses. The total opioid daily doses (Figure 6(A)) decreased significantly (P < 0.05) only at the beginning (day 2) of the intracisternal treatment, increasing thereafter, but remaining under the precisternal values. This was associated with a transient increase of alertness in nine patients, with the ability to read newspapers in five, watch television in nine, perform hand-work in two, and intellectual activity (computerized industrial programming) and return to work in one. The nausea diminished or disappeared in five patients, constipation lessened in five, and respiratory insufficiency decreased in one.

Intracisternal Bupivacaine Daily Doses. The intracisternal bupivacaine daily doses were adapted to the patients' needs, to keep them acceptably free from pain (VAS^mean0-1), and balanced against the occurrence of side effects. This was achieved with hourly doses of 1-7 mg (mean 1.5 mg). Optional bolus doses ranged from 0.5 mg 4 times/h to 5 mg once/h (Table 4). The largest, transient ([nearly equal] 6 h) dose (7 mg/h) was used in a patient with melanoma of the head and neck. Even at this high dose rate, the severe pain provoked by head and neck movements could not be prevented.

Initial and final doses as well as the mean doses during the whole period of treatment are indicated in Figure 6(B). For the whole period, the doses ranged from 20 to 118 mg (median 37 mg) per day. There were significant (P < 0.0001) interindividual differences in the daily doses of bupivacaine but no significant intraindividual variations (P > 0.9) of doses throughout the treatment period.

The cumulative intracisternal bupivacaine doses (Table 4) administered during the period of treatment (3182 days) ranged from 65 to 5,510 mg (median 1,475 mg).

Side Effects and Complications of the Intracisternal Bupivacaine Treatment. The patients were able to eat, drink, speak, walk, and micturate normally, when these functions had been intact before the intracisternal treatment. One patient (a 32-yr-old man) reported sexual intercourse while receiving doses of intracisternal bupivacaine ranging from 2.0 to 3.0 mg/h.

Intracisternal Optional Bolus Doses. Intracisternal optional bolus doses of 1 mg (maximum 4 times/h) showed no adverse effects (Table 5) at a basic rate of 1-5 mg/h. When the bolus doses were increased to 1.5 mg, orthostatic hypotension followed in two patients treated with basic rates of intracisternal bupivacaine of 1.5 mg/h (Table 5). Bolus doses of 5 mg, given to a patient treated with rates of 5 and 7 mg intracisternal bupivacaine per hour made the patient somnolent for approximately 30-45 min. Somnolence was not associated with clinical evidence of seizure activity, hypotension, or hypoxemia. Potential bradycardia could not be detected because the patient was already bradycardic (heart rate 45-60 beats/min) because of treatment with atenolol (Tenormin, Zeneca, Macclesfield, Cheshire, England), a selective beta¹-receptor blocker.

Side effects and complications (Table 5) were related to the intracisternal bupivacaine, as they were reversed when bupivacaine infusion was stopped. They are summarized below.

Motor Side Effects. Signs of motor impairment of the cranial nerves were recorded for the somatic fibers of the vagal nerve, expressed as hoarseness after intrathecal injection of 5.0 mg bupivacaine, in all intracisternal tests, and of the glossopharyngeus nerve (dysphagia) after a bolus dose of 7.5 mg bupivacaine. Paresis of the upper extremities was observed in two patients at intracisternal doses of bupivacaine of 1.5 and 5 mg/h, respectively. No clinical signs of gross phrenic nerve impairment were recorded.

Sensory Side Effects. A feeling of coldness in the neck and skull base (at intracisternal doses of bupivacaine of 2 mg/h) with incomplete phantom pain relief in the amputated upper extremity was recorded in one patient. The catheter was withdrawn [nearly equal] 4 cm. The feeling of coldness in the neck and skull base disappeared, and the phantom pain was completely relieved at the same bupivacaine dose (2 mg/h). Transient paresthesias in the neck and upper and/or lower extremities were recorded in six patients.

Vegetative Side Effects. Episodic miosis and conjunctival hyperemia (as symptoms of Horner's syndrome) were observed in one patient after bolus doses of 1 mg at an infusion rate of 2 mg bupivacaine per hour, after moving the catheter from the intracisternal to the midcervical (C4) position. Episodic, orthostatic arterial hypotension, occurring 1 and 15 days after the start of the intracisternal therapy, was observed in two patients with severe dysphagia and nutritional problems (dehydration and cachexia). The hypotension occurred within 5-10 min after an optional dose of 1.5 mg, given at a basic rate of 1.5 mg/h. Urinary retention occurred in one patient with preexisting urethral stenosis at doses of intracisternal bupivacaine of 5-7 mg/h.

Other Side Effects. Symptoms related to opioid withdrawal were recorded in four patients.

The consequences of sudden withdrawal of intracisternal bupivacaine were observed in one patient treated with 65 mg bupivacaine per day (greater or equal to 2.5 mg/h). The patient previously noted that he had small pupils when the pump was functioning. On the 66th day of the intracisternal treatment, accidental arrest of the pump with interruption of bupivacaine infusion occurred. The patient became restless and agitated. He had tachycardia (100-120 beats/min), dilated pupils, and flushing and profuse sweating in the face and the upper half of the body. The symptoms disappeared within minutes after intracisternal administration of 1 ml 0.25% bupivacaine.

Severe tiredness, faintness, and malaise were recorded in one patient at doses of intracisternal bupivacaine of [nearly equal] 3 mg/h. The symptoms disappeared when the dose of intracisternal bupivacaine was reduced from [nearly equal] 3 to 1 mg/h. Somnolence and sleep occurred in one patient receiving 5-7 mg intracisternal bupivacaine per hour. Repeated, extra doses of 5 mg intracisternal bupivacaine made the patient somnolent for approximately 30 min after each dose, without clinical evidence of seizure activity, respiratory compromise, or cardiovascular disturbances.

Nausea, vomiting, and respiratory depression from the intracisternal bupivacaine were not recorded. Tolerance to intracisternal bupivacaine did not occur, and the individual doses of intracisternal bupivacaine were practically the same throughout the treatment period (Figure 6(B)).

No severe, lasting neurologic sequelae and no death could be attributed to the intracisternal catheterization or to the intracisternal administration of bupivacaine.

The number of observations was too small to permit valid conclusions about the role of continuous intracisternal administration of bupivacaine in the treatment of refractory pain located in the head, neck, and upper limbs, but some notable observations were made.

A relatively stiff nylon catheter was preferred to a soft catheter made of silicone elastomer or polyurethane because of easier insertion, even from a lower cervicothoracic interspace, and maintenance of the catheter tip position at the level of the cisterna magna. However, the stiff nylon catheter should not be used for catheterization of the cisterna magna by the suboccipital approach, because it could damage the medulla and adjacent arteries [17]during insertion and by to-and-fro movements of the tip due to flexion and extension of the neck. This risk probably is nonexistent with a stiff catheter with a closed, rounded tip and lateral holes, inserted from a lower level, because of the straight course of the catheter in the subarachnoid space (Figure 1).

The suboccipital approach [18],*** with a distance between the dura mater and medulla of [nearly equal] 18 mm [19]has the advantage of easier and perhaps safer puncture than a cervicothoracic approach. However, it cannot be used with semistiff nylon catheters, and the catheter cannot be inserted more than 3-4 cm, which will increase the risk of accidental catheter withdrawal. To prevent dislocation, some authors [18]have applied a tissue glue (Histoacryl) around the puncture site. A soft catheter does not keep its position in the cisterna magna because the tip floats and may be displaced downward. Therefore, some authors*** have directed the tip of the Tuohy needle caudally and inserted the catheter only 2-3 cm into the medullary part of the cisterna magna.

A lateral C1-C2 approach was reported in 1994 by Crul et al. [20]in two patients, with the advantage of a lower risk of injury to the spinal cord during advancement of the needle, as it is directed laterally into the anterior subarachnoid cervical space. However, severe complications, such as quadriplegia secondary to hematoma, [21]delayed spinal subarachnoid hematoma, [22]and death [23]have been reported with this approach.

The low cervical/high thoracic (C7-T5) approach [7,8],* has the advantages of a more secure position of the catheter tip at the height of the cisterna magna with the semistiff nylon catheter and a lower risk of accidental catheter withdrawal due to the intrathecal length of the inserted catheter: 24-25 cm in the Ramaioli study* and 14-23 cm in the current study.

The low thoracic (T8-T9) approach was used in two complex cases, when repeated attempts to puncture the dura mater at higher interspaces were unsuccessful. This approach had no other advantages.

The catheter tip was located at the height of the C1 or C2 vertebral bodies, in the lower part of the cisterna magna. With this location, the catheter tip was in the vicinity of the upper posterior cervical roots carrying the sensory fibers, of the trunks of the lower 7 cranial nerves, and of the spinal trigeminal nucleus of the 5th cranial nerve. Other sensory nuclei of the cranial nerves, important for somatic (the main trigeminal nucleus of the 5th nerve) and visceral sensory innervation (nucleus of the solitary tract of the vagus nerve, including afferents from the facial and glossopharyngeal nerves), might be reached by the local anesthetic. Further, with this location, potential injury to the posteroinferior cerebellar artery by the catheter tip during its to-and-fro displacements with flexion/extension of the neck (Figure 1) might be prevented. However, as noted in two patients, the intracisternal location of the catheter tip appeared not to be the best location for treatment of intractable pain from the shoulder and upper limb.

Difficulties encountered during catheterization in the patients with malignant pain may be attributed to pathologic conditions of the spine and spinal canal and the possible presence of epidural metastases.

Intracisternal administration of bolus doses of up to 5 mg bupivacaine appeared to be safe, although with certain manageable and reversible adverse effects in some patients. A positive test might help in excluding centralized pain (e.g., in patients with trigeminal and neck postherpetic neuralgia) before potential intracisternal catheterization.

Continuous intracisternal infusion of bupivacaine in the dosages usually used in this study (1-3 mg/h and 0.5-2 mg as a bolus dose 4-2 times/h, respectively) significantly reduced pain and increased the duration of nocturnal sleep. However, the treatment had no favorable effect on the gait ability pattern scores, because of continuous progression of the basic disease and deterioration of the patients' general condition.

In patients with refractory pain from the shoulders and upper extremities, the intracisternal administration of bupivacaine gave less pain relief than intrathecal administration of the same bupivacaine doses at the midcervical (C4-C5) levels. This adds support to the segmental effects of subarachnoid administration of local anesthetics. [24]The lack of blockade of pain stimuli from other parts of the body (upper extremities included) than the head and neck probably is due to the relatively deep location in the neural substance (with the fibers from the upper limbs deepest) of the spinothalamic tract, carrying the pain afferents from the parts of the body innervated by the spinal nerves. Therefore, the concept that continuous intracisternal administration of bupivacaine (in doses with acceptable side effects) might be a panacea for the treatment of all kinds of refractory pain from all parts of the body is not supported by the current study.

Motor, sensory, vegetative, and other side effects have been presented in detail (Table 5). Paresthesias and pareses in the territories of distribution of the cervicobrachial plexus and the various cranial nerves may be ascribed to the effects of the intracisternally administered bupivacaine. One exception was transient left arm and leg paresis in one patient, for whom edema or formation of a small hematoma from insertion trauma cannot be excluded. Differentiation of sequelae of the basic disease from effects of intracisternal bupivacaine was possible by the reversibility of the symptoms and signs after stopping the intracisternal infusion of the local anesthetic.

The absence of obvious phrenic nerve paralysis is notable and may be explained by the large motor fibers of the phrenic nerve being more resistant than the smaller ones to the effects of local anesthetic agents. [25]Signs of motor impairment of the cranial nerves were found only at bolus doses of 5 mg for the somatic fibers of the vagal nerve (hoarseness) and at 7.5 mg for the glossopharyngeal nerve (dysphagia). No signs of motor impairment from the other cranial nerves were noted at therapeutic doses. These findings are supported by the early observations of the pioneers of high cervical spinal analgesia. [10,11,26].

Of special interest among the side effects of intracisternal bupivacaine is the induction of severe tiredness, faintness, and malaise occurring at doses of greater or equal to 3 mg/h never registered with long-term intrathecal bupivacaine at low cervical, thoracic, and lumbosacral levels. [27]These symptoms may be attributed to a decrease of skeletal muscle tone, inhibition of the tonic activity of the reticular substance by reduction of signals from the peripheral receptors, and loss of control of fairly specific motor tasks, such as rhythmic locomotor movements, and of eye, head, and body movements in response to optic and vestibular stimuli. [28]This might be the mechanism behind the malaise.

Somnolence and sleep may be related to inhibition of the reticular substance by a critical concentration (not known) of bupivacaine in the cerebrospinal fluid surrounding the medulla and brainstem. This may lead to blockade of the reticulothalamocortical pathway, or of the input of impulses to this pathway. The cerebral cortex will be deprived of the stimuli necessary for the maintenance of alertness, which in turn will lead to somnolence and sleep. Such a chain of events finds support in the clinical and electroencephalographic findings of Quatrini et al.**** in their work on "total spinal anesthesia." After intrathecal administration of 270 mg mepivacaine in a volume of 18 ml at the C7-T1 level, these authors noted loss of consciousness within 60-90 s and lasting 60-90 min, diffuse atonia and areflexia, fixed bilateral mydriasis, and electroencephalographic recordings typical of the normal sleep phase 2.

The absence of nausea and vomiting indicates that intracisternal bupivacaine in the doses used in this study has no activating effect, neither on the emetic center in the region of the fasciculus solitarius and the underlying reticular formation nor on the chemoreceptor trigger zone, superficially situated in the floor of the 4th ventricle. This is in contrast to drugs like morphine and apomorphine, which activate the chemoreceptor trigger zone. With intracisternal morphine in doses of 1.5-4.8 mg per day in patients with malignant pain in the head and neck, refractory nausea and vomiting was recorded in 12.5%. [8],* Absence of nausea and vomiting would represent an important advantage of intracisternal bupivacaine when compared to intracisternal morphine. This was our reason for not adding morphine to bupivacaine for intracisternal administration, despite a documented analgetic synergy in somatogenic and viscerogenic pain when these drugs were combined for intrathecal administration. [29,30]We are not aware of any reports on a corresponding synergy in neurogenic or incident pain, in which pain relief might depend on bupivacaine alone and high doses of bupivacaine are sometimes required. [27,31].

The dose of 3 mg intracisternal bupivacaine per hour appeared to be the highest dose for safe intracisternal administration without occurrence of distressing side effects. With intrathecal administration of bupivacaine at lower (lumbosacral, thoracic, and low cervical) levels, the patient will be able to balance the analgesic effect against the side effects from a certain bupivacaine dose. [31]This was not possible with intracisternal bupivacaine treatment. Here, doses above 2 mg/h in patients in poor condition and above 3 mg/h in patients in good physical condition may be associated with unacceptable side effects, such as severe tiredness, somnolence, faintness, and malaise. In addition, 5 or 7 mg intracisternal bupivacaine per hour appeared to be incapable of relieving severe incident pain triggered by movements of the neck and head.

We conclude that long-term intracisternal administration of bupivacaine may help the rare, well selected patient with refractory pain from the head, face, and neck structures to obtain adequate pain relief when all other possible methods have failed. [32]The method probably should not be used in patients with nonmalignant pain when combined with chronic pain behavior and mental disturbances. Further studies are necessary to establish the safety of and indications for the method (e.g., in severe cervical and trigeminal postherpetic neuralgia, when effective pain therapy is otherwise not available). Data from studies of the pharmacologic effects of intracisternal local anesthetics, such as bupivacaine, might increase our understanding of spinal anesthesia, especially the concept of segmental, selective spinal blockade [24,33]in high cervical segments, and extend our knowledge of the effects of subarachnoid local anesthetics at the level of the cranial nerves and the brainstem in humans.

*Ramaioli F, Mara M, Merati P, De Amici D, Ceriana P: Transcutaneous intracisternal analgesia (TIA) (abstract), 10th Annual ESRA Congress, 1991. The International Monitor 1991;32.

**The detailed case reports of individual patients and tables, including additional details of neuropathology, pain treatments before the period of intracisternal infusion, catheterization, intracisternal treatment, and side effects, are on file with the authors and will be sent on request.

***Schoeffler PF, Haberer JP, Monteilard CM. Bouetard ES: Morphine injections in cisterna magna for intractable pain in cancer patients. ANESTHESIOLOGY 1987; 67:A246.

****Quattrini A, Di Bella P, Paggi A, Testasecca D: Valutazione clinico-EEG in soggetti sottoposti ad anesthesia spinale totale, Atti del 1 degree Corso di Aggiornamento sulle Anestesie Periferiche. Edited by Rosano R, Testasecca D. Ancona, Tipografia "Grafiche Bellomo," 1978, pp 338-52.