J Neurosurg 65:32-36, 1986 Percutaneous retrogasserian glycerol rhizotomy for trigeminal neuralgia A prospective study of 100 cases MANUEL J. ARIAS, M.D. Neurosurgical Service, Regional Hospital of Malaga, Mdlaga, Spain ~," A prospective study of percutaneous retrogasserian glycerol rhizotomy (PRGR) with and without metrizamide trigeminal cisternography is reported in the treatment of the trigeminal neuralgia. A series of 100 patients with typical lrigeminal neuralgia were allocated randomly to two treatment groups: Group I patients received PRGR with trigeminal cisternography (50 cases) and Group II patients received PRGR without trigeminal cisternography (50 cases). The results indicate that PRGR without trigeminal cisternography is a valid alternative to the original technique. Factors that assured the accurate performance of the modified technique proposed in this study were: I) spontaneous cerebrospinal fluid drainage; 2) radiologically confirmed placement of the thin spinal needle at the clival edge into the trigeminal impression of the petrous apex and in the center of the foramen ovale; 3) a positive response to the glycerol test; 4) clinical control of the final glycerol injection; and 5) an alert and cooperating patient throughout the entire procedure. KEY WORDS " glycerol 9 rhizotomy ~ trigeminal neuralgia trigeminal cisternography S ZNCE its introduction by Hfikanson in 1981, 3 percutaneous retrogasserian glycerol rhizotomy (PRGR) has become a valuable method for the surgical treatment of trigeminal neuralgia.l'4'5 The objective of this procedure is the percutaneous placement of a small amount of pure sterile glycerol on the intracisternal trigeminal rootlets corresponding to the affected division. To assure accurate intracisternal placement of the needle tip, metrizamide trigeminal cisternography has been advocated as a necessary technical phase of the method. 3'4 However, instead of this radiological confirmation, Sweet, et al., 5 relied on the patient's clinical response to the injection of glycerol in order to identify the trigeminal rootlets to be damaged. As this technical modification seemed to make PRGR safer and simpler, the present study was designed to compare the results of PRGR with and without trigeminal cisternography. Clinical Material and Methods Protocol For inclusion in this study, patients were required to have typical trigeminal neuralgia refractory to or intol- erant of medical therapy. 3 The patients were allocated randomly to one of two treatment groups. Group I patients were treated with PRGR with trigeminal cisternography (50 cases), and Group II patients were treated with PRGR without trigeminal cisternography (50 cases). Patient Population The general characteristics of the patients are delineated in Table 1. Most of the patients were women. The average age was 61 years, and the mean duration of symptoms 5 years. No case of bilateral neuralgia was observed in this study. The pain affected predominantly the lower divisions of the face. No significant difference in general features was found between the two treatment groups. All 100 patients were treated with carbamazepine with an initial favorable effect, which was considered to indicate a typical trigeminal neuralgia. Due to the subsequent ineffectiveness or toxicity of carbamazepine, other drugs (such as phenytoin sodium and baclofen) were used alone and/or in combination. Some patients had undergone surgical procedures prior to this study; peripheral nerve block and neurectomy were 32 J. Neurosurg. / Volume 65 ~July, 1986
Glycerol rhizotomy for trigeminal neuralgia TABLE 1 Clinical summary in 100 cases of trigeminal neuralgia Factor no. of cases 50 100 50 100 100 female 31 62 35 70 66 right side 28 56 26 52 54 age (yrs) < 30 0 0 1 2 1 30-49 7 14 5 10 12 50-69 30 60 32 64 62 70-89 12 24 11 22 23 > 89 1 2 1 2 2 duration of symptoms (yrs) < 1 4 8 6 12 10 1-5 21 42 19 38 40 6-10 22 44 24 48 46 > 10 3 6 1 2 4 affected division 1 3 6 4 8 7 2 9 18 12 24 21 3 8 16 10 20 18 1, 2 7 14 5 10 12 2, 3 20 40 16 32 36 1,2,3 3 6 3 6 6 FIG. 1. Diagram showing the normal variability of the shape and size of trigeminal cisterns and their common location (A) by the middle third of the clival edge immediately before the trigeminal porus (B). The needle (C) must be directed to this point. predominant among these. Two patients in both groups received radiofrequency thermocoagulation, one patient in Group I was treated by retrogasserian rhizotomy by the subtemporal route, and one in Group II by microvascular decompression via the posterior fossa. Preoperative Management In addition to a careful neurological study with special attention to facial sensation, x-ray films of the skull (anteroposterior, lateral, and submentovertex projections) and a computerized tomography (CT) scan were obtained in all patients. Most of the cases were operated on as outpatients. The night before the operation, the patients received 15 mg clorazepate by mouth, and 1 hour before the operation 40 mg of metoclopramide and 0.5 mg of atropine sulfate were injected intramuscularly. No patient received anesthesia during the operation, and all were alert and cooperating throughout the entire procedure. Operative Technique All operations were performed in the neuroradiological room with the patient on a radiological table convertible to a chair, with a G-arm attached for fluoroscopy and roentgenograms. The materials used included disposable 1-ml syringes, a thin disposable No. 22 lumbar spinal needle, metrizamide (300 mg/ml iodine), and pure sterile glycerol. Facial Puncture. The initial placement of the needle by the anterior route was performed with the patient in the supine position on the radiological table. The needle was inserted through the cheek 3.5 cm from the corner of the mouth and 0.5 cm below; it was directed to a point 3 cm anterior to the external auditory canal and at the medial aspect of the pupil. With the aid of intermittent lateral fluoroscopy, the puncture of the foramen ovale was easy and painless in most of the cases. Verification of Needle Position. Correct localization of the needle tip in the trigeminal cistern was verified by: 1) the appearance of spontaneous cerebrospinal fluid (CSF) drainage; 2) radiological evidence; 3) metrizamide trigeminal cisternography; and/or 4) a glycerol test. Spontaneous CSF drainage was considered a favorable sign of intracisternal placement of the needle tip and a prerequisite to continuing the procedure except when it was suspected that the cistern was obliterated by prior surgical gasserian procedures. However, spontaneous CSF drainage may also occur if the needle tip is located in the subtemporal subarachnoid space. Radiological evidence was also used to verify a correct location of the needle. In the anteroposterior projection, the needle tip must lie in the trigeminal impression of the petrous apex. If the needle appears to be too laterally located in this view, it may have penetrated into the subarachnoid space. In the lateral view, the needle tip must be located at the middle third of the clival edge. Although the distance between the foramen ovale and the trigeminal cistern does vary, there is a region around the clival edge where almost all trigeminal cisterns reach, and it is to this point that the needle tip must be directed (Fig. 1). In the submentovertex view, the needle must be visible at the center of the foramen ovale. When it is located laterally it is probably extracisternal or projecting into the subdural space of J. Neurosurg. / Volume 65 ~July, 1986 33
M. J. Arias TABLE 2 Final results in 100 patients with trigeminal neuralgia Factor* no. of cases 50 I00 50 100 100 complete relief 47 94 48 96 95 immediate 45 90 44 88 89 late ( 1 st wk) 2 4 4 8 6 failures 3 6 2 4 5 no brisk egress of CSF 2 2 4 negative glycerol test 3 2 5 prior gasserian operation 2 2 4 age < 50 yrs 2 1 3 recurrences 6-12 mos 0/13 1/14 1/27 13-24 mos 1/11 0/12 1/23 25-36 mos 4/26 4/24 8/50 * CSF = cerebrospinal fluid. Recurrences are expressed as no./no. of cases at risk. Meckel's cave, and it is advisable to reposition the needle before continuing. After the correct position of the needle tip had been verified by spontaneous CSF drainage and radiological control, metrizamide trigeminal cisternography was performed in the 50 patients randomly allotted to Group I. The technique described by H~kanson was followed. -''3 No attempt was made to completely fill the cistern with metrizamide because this technique is used only for the identification of the cistern. The glycerol test was applied to all patients. This involves the injection of 0.05 ml of pure sterile glycerol into the trigeminal cistern with the patient in the sitting position and the head flexed depending on the trigeminal division affected (about 40 ~ for the first division, 25 ~ for the second, and almost erect for the third). After glycerol injection, the patient's facial sensation is tested. The response usually occurs between 1 and 5 minutes following injection. A response is considered positive when the patient feels focal paresthesias (such as cold, warmth, numbness, tingling, prickling, crawling, compression, itching) on one or more trigeminal divisions on the side of the face to be injected, usually close to the midline around the mouth. Another positive response, although rare, is when a complete or partial attack of neuralgia is produced. The glycerol test demonstrates not only the intracisternal location of the needle tip but also the correct degree of flexion of the head to facilitate selective damage to the rootlets corresponding to the affected trigeminal division. By slowly changing the degree of head flexion, the response to the glycerol test can be transferred from one trigeminal division to the next. Glycerol Injection. Finally, the glycerol injection was performed with the patient in the position determined by the best response to the glycerol test. Glycerol was slowly injected in small 0.05-ml increments. After each increment an interval of about 5 minutes was allowed during which the symptoms and the sensation of the face were tested. Glycerol injection was stopped when the first signs of facial sensory loss were detected or when the maximum volume of reference (first division: 0.10 ml, second and third divisions and first to second divisions: 0.25 ml; second to third divisions: 0.30 ml; and first to third divisions: 0.40 ml) was reached. After the final glycerol injection, the patient usually remained in the same position for an additional hour, except when the facial sensory loss was too great (> 30% loss). All patients were tested for facial sensation before, during, and after the procedure with pinprick and cotton wool. During the operation the assessment was made after the glycerol test, after each increment of the glycerol injection, at the end of the operation, and 1 hour later. Postoperative Management All the outpatients were discharged 2 hours after the operation. The inpatients remained hospitalized for 24 hours. Prior medical therapy was tapered gradually over 2 weeks starting 15 days after the operation. Sensory examination was performed at the end of the 1st, 4th, and 9th weeks. Results Forty-seven patients (94%) in Group I and fortyeight (96%) in Group II had complete relief of their trigeminal neuralgia and no longer needed medical therapy (Table 2). Two of the 47 patients in Group I and four of the 48 in Group II experienced relief after a delay of 2 to 4 days. Three (6%) patients in Group I and two (4%) in Group II failed to gain relief. Retrospective study showed that common factors in these failures were: age under 50 years, no brisk egress of CSF, a negative glycerol test, and a prior gasserian operation. At follow-up review (range 6 to 36 months) five recurrences were found in each group; most of these occurred more than 2 years after operation. In Group I a second operation was performed in five cases and a third in four; in Group II a second operation was conducted in four cases and a third in two. Therefore, 59 operations were needed in Group I to relieve 47 patients, with three failures and five recurrences; in Group II, 56 operations were required to relieve 48 patients, with two failures and five recurrences. During the operation, nausea and/or vomiting were observed in only two patients, both in Group I (Table 3). In the patients with basal headaches, onset occurred when the contrast medium and/or the glycerol were injected extracisternally. Most patients had paresthesias during the glycerol test and only a few had an attack of neuralgia. Facial sensory loss in the affected division was minor in most of the patients at the end of the operation; in a few cases it included the adjacent division. Reduced corneal sensation of a minor degree (< 50% loss) was found in only two patients. Two pa- 34 J. Neurosurg. / Volume 65 ~July, 1986
Glycerol rhizotomy for trigeminal neuralgia TABLE 3 Treatment features in 100 patients with trigeminal neuralgia Factor no. of cases 50 100 50 100 100 intraoperative features 2 4 0 0 2 nausea &/or vomiting 14 28 4 8 18 basal headaches 46 92 48 96 94 positive response to glycerol test 40 80 44 84 84 focal facial paresthesias 6 12 4 8 10 neuralgic attack facial sensory loss affected division: 30%-60% loss 1 2 1 2 2 affected division: < 30% loss 34 68 35 70 69 adjacent division: < 30% loss 7 14 10 20 17 reduced corneal sensation 1 2 1 2 2 postoperative complications (1 st wk) nausea &/or vomiting 2 4 0 0 2 transient aseptic meningitis 2 4 0 0 2 basal headaches 8 16 1 2 9 herpes simplex perioralis 4 8 6 12 10 facial sensory loss affected division: 30%-60% loss 0 0 0 0 0 affected division: < 30% loss 24 48 25 50 49 adjacent division: < 30% loss 3 6 5 10 8 reduced corneal sensation 1 2 1 2 2 tients continued to vomit during the 1st postoperative week; they had a transient aseptic meningitis which responded promptly to corticosteroid therapy. Ipsilateral eruptions of herpes simplex perioralis developed in 10 cases. Facial sensory loss decreased significantly by the end of the 1st postoperative week. Table 4 delineates the evolution of the facial sensory loss from operation to the 90th postoperative day when only 13 patients remained with a minor sensory loss of the affected division and no case presented with loss in an adjacent division or the cornea. No serious complications were observed in this study. No patient developed keratitis, anesthesia dolorosa, masseter weakness, diplopia, or vascular damage. Discussion Final results of this study corroborate previous similar reports demonstrating that PRGR is an effective, innocuous, and simple method for the surgical treatment of trigeminal neuralgia. ~'3-5 Ninety-one patients achieved complete pain relief after the first injection, three after the second, and one after the third. Thus, 95% of the patients were free of pain after one or more operations. There was no significant difference between the results in the two treatment groups. The five treatment failures in this study presented some common features from which it is reasonable to infer that extracisternal injection and/or trigemina[ cistern obliteration negatively affect the results. As the final injected volume of glycerol in our technique depends more on the statement of the patient as to facial sensation than on TABLE 4 Evolution of facial sensory loss Facial Group I Group ll Total Senso~ Loss affected division at operation 35 70 36 72 71 8th postop day 24 48 25 50 49 30th postop day 8 16 7 14 15 90th postop day 7 14 6 12 13 adjacent division at operation 7 14 10 20 17 8th postop day 3 6 5 10 8 30th postop day 1 2 2 4 3 90th postop day reduced corneal sensation 0 0 0 0 0 at operation l 2 1 2 2 8th postop day 1 2 1 2 2 30th postop day 0 0 0 0 0 the volume of the cistern, insufficient volume of glycerol does not appear to be a factor influencing the final results. It is not necessary to know the total volume of the trigeminal cistern if the goal is to selectively damage only a group of trigeminal rootlets. It is unclear from the report of Lunsford and Bennett 4 how to deduce the total volume of the cistern or the amount of glycerol to be injected. On the other hand, neuralgia recurred in only 10% of the cases in the present study: five cases in each group. These recurrences tended to present more than 2 years after the operation and had no relation to the number of operations needed to relieve the pain. As in similar reports, this study included no cases of anesthesia dolorosa, keratitis, masseter weakness, diplopia, vascular damage, or other serious complication. 3-5 In our series, facial sensory loss was of minor degree and progressively faded, and there was no significant facial loss during or immediately after the glycerol injection as was reported by Sweet, et al) This feature could have been due to an excessive amount of glycerol and/ or to an intraganglionic injection. 4 Intraganglionic injection of the glycerol is more likely to occur when the needle tip is too close to the floor of the middle cranial fossa because the gasserian ganglion is always closer to the foramen ovale than the trigeminal cistern. Glycerol is a weak neurolytic agent that provides a degree of deafferentation depending upon the volume injected, the duration of application, and individual response to its chemical composition. One of the better and simpler ways to measure this degree of deafferenration is by assessing continuously the clinical response of the alert patient during the glycerol injection. Using this technique, it is possible to use a minimum volume of glycerol to selectively damage the rootlets involved. With the technique used in this study, it is possible to "move" the glycerol intracisternally from one group of rootlets to another by modifying the degree of flexion of the head. In this sense, the glycerol test can be J. Neurosurg. / Volume 65 ~July, 1986 35
M. J. Arias compared with the electrical stimulation technique of radiofrequency thermocoagulation to localize the rootlets to be damaged. Our results show no significant difference between the two experimental groups. Aseptic meningitis and basal headaches predominated in Group I, the group with trigeminal cisternography. The use in the same procedure of two chemical substances injected into the basal cisterns could increase their toxic effects on the meninges. It is important to emphasize that final results and facial sensation were similar in both groups, so that PRGR without trigeminal cisternography was, in this study, as effective and innocuous as with it. Also, PRGR without trigeminal cisternography is simpler, cheaper, and safer. The identification of the trigeminal cistern is only a means to verify the trigeminal rootlets that correspond to the trigeminal division affected, so as to assure accurate placement of the glycerol. Even after trigeminal cisternography, the glycerol injection can be inadvertently extracisternal if the needle tip is displaced during the maneuvers of the injection -- a circumstance that could escape notice without a continuous check on the facial sensation of the alert patient. This study demonstrates that the glycerol test and clinical control of the glycerol injection are technical modifications that improve the original technique of PRGR. References 1. Arias M J: Tratamiento de la neuralgia trigeminal mediante rizetomia percutfinea retrogasseriana con glicerina. Rev Neurol (Barcelona) 12:219-224, 1984 2. H~kanson S: Transoval trigeminal cisternography. Surg Neurol 10:137-144, 1978 3. H~kanson S: Trigeminal neuralgia treated by the injection of glycerol into the trigeminal cistern. Neurosurgery 9: 638-646, 1981 4. Lunsford LD, Bennett MH: Percutaneous retrogasserian glycerol rhizotomy for tic douloureux: Part 1. Technique and results in 112 patients. Neurosurgery 14:424-430, 1984 5. Sweet WH, Poletti CE, Macon JB: Treatment of trigeminal neuralgia and other facial pains by retrogasserian injection of glycerol. Neurosurgery 9:647-653, 1981 Manuscript received October 8, 1985. Accepted in final form February 4, 1986. Address reprint requests to." Manuel J. Arias, M.D., Marbella I0, Fuengirola, M~tlaga, Spain. 36 J. Neurosurg. / Volume 65 ~July, 1986