Microvascular Decompression; Abnormalities at the Trigeminal Nerve Root Entry Zone

A steadily increasing number of neurosurgeons are so favorably impressed with the results of microvascular decompression (MVD) of the trigeminal root entry zone at the pons in the treatment of trigeminal neuralgia that they recommend this for all but medically infirm or older patients when medical management becomes unsatisfactory. The principal uncertainty in the minds of some of us is the extent of the pressure on the trigeminal rootlets required to provoke the syndrome. In some cases causative pressure by one or more arteries exists. Jannetta described "neurovascular compression" at the root entry zone in 395 (96 percent) of 411 patients with trigeminal neuralgia. Zorman and Wilson stated that "veins and arterial loops not in direct contact with the root were presumed to be causative if they were in close proximity to an anatomically deformed nerve root because retracting the cerebellum could displace the pons and trigeminal root away from an offending artery or vein. They found causative extrinsic lesions in 79 percent of 125 patients. Their results after MVD were about the same whether they treated presumed venous or arterial compression or, finding neither, did a partial sensory rhizotomy (PSR). Breeze and Ignelzi likewise considered the finding of a vessel "near" the nerve in 37 percent of their 52 procedures as indicative of the cause of the pain.

In a 1989 review of 252 patients with trigeminal neuralgia treated between 1969 and 1985, Wilson's statements seem to reflect a striking change in his criteria for determining treatment of significant compression of nerve by blood vessel. When the nerve root is "distorted" by an artery and/or a vein, he does an MVD by coagulating and dividing veins, and by dissecting arterial loops from the nerve root. This displacement is maintained by a small sculpted prosthesis of sponge Teflon inserted between the root and the vessel. One drop of cyanoacrylate cement is used to secure the prosthesis to the pia or nerve root. Patients in whom vascular contact was found without deformity of the nerve root underwent both displacement of the vessel (MVD) and PSR consisting of section of the inferior one-half to two-thirds of the portio major. Patients in whom no vascular contact was found underwent PSR only. Using these criteria he performed MVD only in 166 (66 percent) of the cases, MVD and PSR in 56 (22 percent), and PSR only in 30 (12 percent). Seven local tumors and two arteriovenous malformations (AVMs) were included in the group undergoing MVD and no PSR. In 1984, Wilson published a review of 118 of the earliest cases of this series and reported that 92 (78 percent) underwent MVD only while 26 (22 percent) had a PSR; none had both procedures. In this earlier series the therapeutic results in the two groups were indistinguishable. In the same year, however, Piatt and Wilkins described a drop in the success rate of MVD from 83 percent to 62 percent when the artery and nerve were merely in contact. Wilson was apparently impressed by this observation, because in his later publication he added PSR to MVD in the management of this arterial contact group. The resultant percentages in the "excellent" and "good" categories were slightly better in these patients than in those who had only one of the two procedures. In these latter two groups the outcomes were about the same. With respect to complications, Wilson achieved a major triumph, with no deaths and no lasting deficit of any significant function in any patient. In seeking to match his surgical judgment and dexterity by his mental agility, Wilson argued that posterior fossa exploration should be the first­line therapy, stating that even successful prolonged "medical treatment may ultimately reduce the efficacy of surgery." He and Alksne have drawn attention to their finding that any proximal ablative procedure, including any percutaneous tactic in the middle cranial fossa, decreases the likelihood of success of a subsequent MVD. But, as Jannetta (1985) asks, "are these patients just less amenable to any therapy?" It is obvious that ease of controlling the paroxysms of trigeminal neuralgia varies enormously from case to case. In Wilson's series a second posterior fossa procedure was performed in 20 patients with persistent or recurrent pain. In no patient had the prosthesis shifted so it was left in place. He logically did a PSR in 18 of the 20 patients, 7 of whom had undergone a PSR at the first operation, achieving an excellent or good result in 85 percent. It seem that Wilson had general shift to requiring more reasons for doing a PSR that he is less confident that extrinsic vascular compression causes most of the cases of "essential" trigeminal neuralgia.

Wilson also pointed out that his' 'excellent" results were in patients who had a shorter duration of symptoms (average 7.4 years) than those in the "good," "recurrent," and "poor" categories; accordingly, in his view we should operate early "before the nerve is irreversibly damaged." However, his "good" category's average duration of symptoms of 10.5 years is much longer than the worse "recurrent" and "poor" categories. This argument for early operation is made the flimsier by his finding that the younger the patient at the onset of the symptoms, the worse the result of MVD. Hence in those 20 who failed to maintain benefit from sustained adequate decompression as well as in the 30 who never had any evidence of extrinsic compression, totaling 20 percent of the 252 cases, such compression did not cause their pain.

The French neurosurgeon Sindou and his colleagues are the only ones who have compared the cerebellopontine (CP) angle findings at operation between patients with trigeminal neuralgia and those with facial pain due to neoplasm. Among 150 patients in the former group they found a vasculoneural "conflit" in 132 (88 percent) of them. A further 14 patients (9 percent) had some other obvious lesion. However, they found a vasculoneural "conflit" in not a single one of 52 patients in whom they were doing a trigeminal rhizotomy for pain due to tumor located away from the sensory root.

Sindou and colleagues have thoughtfully attempted to determine whether the pain-relieving feature of "microvascular decompression" is really the removal of a compressive lesion-the vein or pulsating artery- or is merely another form of compression-a "neocompression." Does an interposed substance-muscle or other material-between the nerve and vessel simply shift from a vascular to another form of constant compression of the rootlets, analogous to the temporary frank rootlet compression of the open or percutaneous operations in the middle cranial fossa. Accordingly they studied decompression by two relevant differing methods. In two clinically similar groups, each of 60 patients with trigeminal neuralgia, the first group had the artery dissected away from the nerve and maintained in that position by the interposition of foreign material (Dacron) and at times some periostium as well between the artery and nerve. In the second group the offending vessels were dislodged without any material touching the nerve. The artery was held away from the nerve by Teflon strips, at times with a 7 x 10 mm piece of Dacron also. A follow-up at 1 year revealed a recurrence rate of 10 percent in the first group and 4.5 percent in the second. They concluded that a "true decompression" rather than a "neocompression" was the more effective type of operation.

Electron microscopic study of the rootlets distorted by the blood vessel and of other specimens maximally removed from the site surely is the most direct way to secure relevant data as to the cause of the paroxysms.

At the other end of the scale are the observations of Adams et al., who saw "Indentation, grooving, or distortion of the nerve by a vascular structure" in only 6 of 57 patients explored in the posterior fossa. Similarly Rovit noted such finds in only 2 of 25 patients.

Postmortem studies to determine the neurovascular relationships in patients without facial pain are about as divergent as the in vivo observations on those with trigeminal neuralgia. The four publications to evaluate are those of Hardy and Rhoton (1978), Haines et al. (1980), Mehta et al. (1981), and Klun and Prestor (1986 ). The first group saw arterial contacts in 52 percent of 50 CP angles, "often" with arterial compression. In contrast were the observations of Mehta et al., who saw arterial contacts with the trigeminal root in only 13 percent of 60 CP angles, and no distortion of nerve by artery. The following were the figures in non-tic cadavers of Haines et al. in 40 CP angles: 25 percent arterial contact, 23 percent venous contact, and 10 percent compression. The corresponding figures of Klun and Prestor in 130 CP angles were 23 percent, 4 percent, and 7 percent.

In view of the far higher percentages in the cadavers of contacts, arterial or venous, than of compression recorded by all four groups, it clearly makes sense to correlate the percentages achieving pain relief with the various types of neurovascular relationship. The first ones to do this. Piatt and Wilkins, found that when an artery distorted the root or was wedged between the root and the pons, as seen in 37 cases, MVD realized an expected cure rate of 83 percent. This dropped to 62 percent of 31 cases when the artery to rootlet relation was one only of contact. A further drop to 42 percent occurred in those with merely venous contact or distortion, or "no abnormality." Their highest percentage of excellent and good results ensued in their 21 patients in this last status on whom they did a PSR.

Comparison of Results of Minimally Destructive Operations in the Middle or Posterior Fossa

There is good evidence that there is some other causative factor in the majority of cases not related to the anatomy of the root entry zone at the pons or the nearby root beyond that zone. These begin with the results of the various largely nondestructive operations in the middle cranial fossa carried out in the fifties and sixties in an effort to avoid the deafferentation dysesthesia (anesthesia dolorosa) of trigeminal root section. A variety of procedures were performed. Taarnhoj stayed with his proposal of "decompression" by incising the length of the dural roof of the trigeminal cistern. Shelden et al. noted that their original concept of "decompressing" the peripheral second and third divisions at the foramina rotundum and ovale gave about the same results as the Taarnhoj operation. They decided that the feature common to the two maneuvers was probably modest trauma to the trigeminal pathways; hence, they changed the decompression to "compression," rubbing vigorously the ganglion and rootlets with a blunt instrument after opening the dura over them. Several minor variations on these concepts were proposed by other neurosurgeons. Thus Gardner "frees the sensory rootlets from their dural sleeve, gently touches them with a cotton pledget, and sprays them forcefully with a Ringer's solution." Other modifications of local manipulation by six other neurosurgeons (summarized by White and Sweet) all shared the feature of cutting no rootlets. Over 1600 cases thus treated had been reported by 1969.

TABLE-1 Results of Minor Manipulations in the Middle Fossa
Reference	Number of Patients	Initial Failure+ Significant Recurrence (%)		Duration of Follow-up (Years)
Bohm & Hojeberg	111	24	22%	2-4
Gardner & Miklos	102	25	25%	3.5-5.5
Graf	100	26	26%	Up to 8
Hamby	88	7	8%	Up to 6
Rowe	96	14	15%
Shelden	200	50	25%	Up to 10
Taarnhoj	230	93	40%	Up to 20
				Avg >12
Totals	927	239	26%

Table-2 gives similar data from 17 series of patients treated by MVD. The duration of follow-up is roughly similar in the two groups except for the cases of Taarnhoj that were traced much longer for an average of more than 12 years. If these are deleted from Table-1, the percentage of initial failure plus significant recurrence for the middle fossa operations drops from 26 to 20.9 percent.

Several reports describe the great majority of recurrences as taking place within the first year or two both for the middle fossa manipulations and for MVD in the posterior fossa. Tables-3 and 4 indicate from the meager long-term data available that further late recurrences in fact continue to appear. Although after both types of operation there may be a period of some days to a week before the neuralgic pains subside, they did so after the middle fossa procedures in all but 24 of 637 cases

percent) reported by 14 surgeons. After MVD, initial sustained failure occurred in 45 of 668 (6.7 percent) in eight reports.

Dahle et al. in their 57 cases found that the procedures failed to give complete relief in 44 percent of their patients (average follow­up only 3.1 years).21 Their results included one fatal cardiac infarc­tion at I week; recoveries from a hemiparesis and partial bulbar palsy at 3 months, and from bacterial meningitis in two cases; and two permanent sequelae-a patient with severe hemifacial sensory loss and another with unilateral deafness and disabling dysesthesia. That these surgeons should continue with MVD op­erations in the face of the data on percutaneous operations is in­teresting.

A Japanese group, Yamaki and coworkers, graded the degree of neurovascular compression as + for definite nerve grooving or

Long-Term Results of Microvascular Decompression
	Number of Patients	Initial Failure (%)		Late   Recurrence (%)		Total (%)		Duration of Follow-up
Apfe1baum4	300	18	(6)	57	(19)	75	(25)	Avg., 63 months (up to	10 years)
Barba & Alksne6	37					8	(22)	Avg.. 42 months
Breeze & Ignelzi 14	52	8	(15)	7	(13)	15	(29)	Avg., 23 months
Dahle2 I	57	14	(25)	11	(19)	25	(44)	Avg.. 3.1 years
Ferguson et a!. 00	24	3	(13)	2	(8)	5	(21)	Avg.. 28 months
Jannetta41	411					81	(20)	1-10 years
KIun4J
MVD	178	4	(4)	11	(6)	18	(10)	Avg.. 5.2 years
PSR*	42	6	(14)	22	(49)	28	(63)	Avg.. 5.2 years
K.olluri & Heros45	72	3	(4)	13	(18)	16	(22)	Avg., 5.9 years
Panagopouloso9	30	2	(7)	4	(13)	6	(20)	6 months-8.5 years
Sengupta
Pen)' et a!. 62	50	2	(4)	2	(4)	4	(8)	4 months-6 years
Pian & Wilkins6J	81					21	(26)	Avg., 48 months
Richards et al.67	25	1	(4)	0	(0)	I	(4)	Avg.. 30 months
Sindou et a!. 76	113	8	(7)	9	(8)	17	(15)	1-6 years
Taarnh0jHO	86					17	(20)	Up to II years
Voorhies & Patterson84	32	2	(6)	3	(9)	5	(16)	Avg.. 23 months
YamakiH7	60
No relief		4	(7)	9	(15)	13	(22)	0.5-5.5 yrs
Partial relief		4	(7)	13	(22)	17	(28)
an & Wilson90	92					II	(12)	Avg.. 26 months
	1700	73/993	(7)	141/993	(14)	383/1700	(22)
-third or less of portio major.
e: Adapted from Gybels and Swcc!..'"

distortion, :!:: for probable arterial contact, and 0 for none (Table 406-5).87 Their early and late failures included major pains in 29 percent and partial recurrence in 35 percent.

Results of Retrogasserian Glycerol Injections

The procedure, pioneered by Hitkanson, of injecting glycerol into the fluid of the tiny trigeminal cistern has found favor in many quarters because of the minimal numbers of patients developing dysesthesias and showing any lasting significant trigeminal sen­sory loss after such injections. Once the small quantities of glyc­erol injected leave the cistern they are diluted by the much larger volumes of CSF in the CP and cerebellomedullary cisterns. Hence, this toxic effect must be confined to that portion of the rootlets in the middle cranial fossa. Table 406-6 summarizes the findings in 22 services on many of which the results have been found so satis­factory that this is their invasive procedure of first choice when the medical management of trigeminal neuralgia has failed. 5.7.10, 19,22,23,26,29,36,37.40.48,49,50,57,65,70.71 ,77.83,88 Although the percentages of early and late failures exceed those of either of the two open operations discussed, the modest degree of sensory loss may be comparable. There is essentially no mechanical manipula­tion of the rootlets involved and this may account for the less favorable results. 'As with the other procedures, the longer the follow-up the higher the recurrence rate,

The additional eight publications tabulated in this edition (Table 406-6) reveal even higher recurrence rates, making this drawback progressively more definite.1O,22,26,29,40,57.70,83 In the papers of Fujimaki et al. and of North et aI., Kaplan-Meier curves are plotted, revealing a recurrence rate at 3 years of 52 and 50 percent, respectively, and at 4'/2 years of 72 percent in the paper of

TABLE 406-4 Late Results of Microvascular Decompression in the Posterior Cranial Fossa

			Late
	Number	Initial	Major	Duration
	of	Failure	Recurrence	of
Reference	Patients	(%)	(%)	Follow-up
Burchiel et al?O	36	10	28*	Avg" 8.5 years
				(7,5-11.5
				years)
Pollack et al. 64	35t	11	31	Avg" 6,3 years
				(1- > 96
				months)

Fujimaki et a1.29,57 In the other four papers the total failure rates were 29, 40, 45, and 71 percent. De la Porte et al. state only that the rate of pain relief was lower than that with thermocoagulation or percutaneous compression.22

A point I have noted, the tendency of the glycerol to strike VI fibers more often than those in the second and third divisions, has been much more quantitatively shown by Bergenheim, Hariz, and Laitinen.lO They measured the thresholds for facial pain perception in milliamperes at six standard sites (using the ISSAL 1412 appa­ratus), delivering a constant current to bipolar electrodes. The day after the glycerol instillation into the trigeminal cistern in 57 pa­tients the threshold cU/Tent for perception of any sensation and the higher threshold for pain were increased as compared with preop­erative thresholds. They say this increase "was especially marked and well localized for those with pain in the first branch." (This loss need not extend to corneal anesthesia.) The absence of any sensory loss and minor sensory loss in the remainder characterized the first 100 patients of Hitkanson and, as cited in Table 406-6, preservation of much or all sensation was the finding of most, but not all, of the 22 rapporteurs. There is essentially no mechanical manipulation of the rootlets or increased pressure against them and this may account for the less favorable results. As with the other procedures, the longer the follow-up the higher the recurrence rate. However, the fact that the majority have long-lasting relief associ­ated with minimal sensory loss adds another large group of pa­tients to those whose relief cannot be attributed to any change in neurovascular relations at the trigeminal root entry zone into the pons.

Conclusion upon Comparison of Open Operations

The similarity between the general results in the two groups of open operations is striking. The likelihood of modification by mid­dle fossa manipulations of anatomic relationships at the pontine trigeminal root entry zone is so small that the hypothesis that local pressure there causes the pain is for the great majority of cases untenable by virtue of this set of facts alone. Adams, in a com­mendable detailed analysis of many factors bearing on the hypoth­esis, has adduced an impressive array of considerations leading to the same conclusion,2 none of which alone seems to me as telling a refutation of the hypothesis as the results of minimally destructive middle fossa operations. However, taken in toto they add up to a major challenge to, but do not disprove, the concept of vascular compression as the most common cause of trigeminal neuralgia. Skepticism as to the validity of this concept comes as a bitter disappointment to many neurosurgeons, including me. The micro­neurosurgery involved in "vascular decompression" is such a de­lightful exercise that it saddens me to recommend great reduction in its usage because its mortality and morbidity cannot be reduced to the levels of the percutaneous operations.

Other Evidence of the Salutary Effect of Manipulation Near the Nerve

	+	:t	0
Cure	26	5	2
Recurrence	8	3	0
Minor
Major	7
Total	41	9	3
* Seven of 60 not classified.
Source: Data from Yamaki et al.87

Operative procedure

Mullan's thoughtfully developed tactic is to pass a 14-gauge nee­dle to but not through the foramen ovale and to inflate a Fogarty catheter's balloon placed into Meckel's cave to about 10 mm be­hind the foramen ovale. This has proved to be a rational major improvement on the results of the open operation and to have im­portant advantages over the other two percutaneous procedures. With the patient under general anesthesia the balloon is inflated with 0.5 to 1.0 ml of a water-soluble contrast medium and ob­served fluoroscopically to make sure the expansion is primarily in * posterior part of the middle cranial fossa, preferably with a 8D3l1 posterior nubbin bulging through the porus trigemini into the

rior fossa. Frequent observation is needed to detect promptly ) a sudden disruption of the dural sheath of the ganglion with enlargement of the balloon, (2) migration of the balloon into the posterior fossa, or (3) rupture of the balloon. None of these prob­lems caused a lasting sequel. For example, Fraioli's balloons broke six times without causing any complication.26 However, in Lobato's four cases, when the balloon moved into the posterior fossa for 15 to 30 sec a diplopia lasting I week to 4 months en­sued.47 Abdennebi and Amber, having cut two balloons with the sharp edge of a trocar or needle, recommend enveloping the metal with a Teflon catheter.

Lichtor and Mullan found that 1 min of inflation sufficed to control the pain and reduced the incidence of persistent dysesthesia.

Pressure in the Balloon: Degree and Duration

There is a little uncertainty as to the proper degree and duration of balloon pressure. Lobato et al. infrequently exceeded I min.47 As noted, Lichtor and Mullan decided after 60 cases to drop the dura­tion to I min.46 Likewise, Fraioli et al. discontinued their 3- to 7-min range in favor of I to 2 min of inflation.26 The reverse tactic has been adopted by the Abdennebi and Meglio groupS.1,52.53 The former tried initially a range from 4 to 9 min for 20 cases, adopting the 7- to 9-min range for the last 30. The most striking differences between short and longer compressions were recorded by Meglio et al. with only 76 percent of 21 patients securing immediate pain relief and 84 percent experiencing recurrence after balloon com­pression of I to 3 min. 52 All 53 of their patients undergoing com­pression for 4 to 10 min had immediate relief and in only 57 per­cent had the pain recurred. Their high recurrence rates are in some

measure explained by the inclusion therein of minor medically controllable pains.

The logical approach of measuring the intraballoon pressure has been taken by three groups, especially when it became appar­ent that there are substantial variations in the force required to distend the balloons. Zanusso's 22 patients were classified in three groupS.89 The three patients with pressures of 0.9 to 1.3 bars had only temporary relief of pain and no sequelae; the 11 at 1.4 to 1.8 bars were all initially relieved, but I had a recurrence and 4 had

sequelae; the 8 at 1.9 to 2.4 bars all remained relieved, but all had sequelae (I bar = 760 mmHg). Lobato's group measured this pressure in 22 of their 144 cases, correlating the appearance of the lateral radiographs therewith.47 Intraluminal pressures of 700, 1150, and 1700 mmHg were seen with, respectively, (a) no bulge toward the posterior fossa, (b) a bulge with a pear shape, i.e., a posterior nipple, and (c) a marked bulge into posterior fossa pro­ducing "intense hemifacial numbness." Brown and Preul found that the recommended pear shape developed at a mean of

815 mmHg, but this ranged from 459 to 1273 mmHg.17.18 Lobato et al. say that pressures under 600 mmHg are going to fail.47 Clearly more data are needed to determine whether or not these measurements can teach us to improve the precision of the opera­tion. Lobato relied on the degree to which the balloon adopts a "fully developed" pear shape, stating that this configuration "al­most invariably resulted in long-lasting control of neuralgia." Mullan and Lichtor, placing emphasis on avoidance of dysesthesia, prefer to distend only' 'until the balloon begins to assume a pear

shape," . 'indicating that it is beginning to protrude out of the cave toward the posterior fossa. ,,46.56

Results of Others

Emphasizing their endorsement of this procedure. three other groups, led by Fraioli, Lobato. and Meglio. have each published a second paper pointing out continued satisfaction with longer follow-ups. In particular. the incidence of facial dysesthesia severe enough to require medical therapy is less than nearly all of us have reported in large series of percutaneous thermal rhizotomies. The brevity and painlessness of the procedure under continuous general anesthesia and the simplicity of the technique are all important advantages. The results of the 12 principal groups reporting are given in Table 406-81.9. ]:i.] 7.] 8.24.26.27A6A7.51.52.61.8-'.89

Multiple Sclerosis

The Lichtor and Mullan cases include five patients with multiple sclerosis. three of them with bilateral disease.46 Of the seven sides treated by PTC there were failures on only two sides. On one of those sides two PTCs failed. each after light compression because of dense numbness produced by PTC on the first side. Fraioli et al. gave early relief in one of three patients but the pain recurred.26 Multiple sclerosis is also more difficult to treat by thermal rhizot­omy than is "essential" trigeminal neuralgia.

Repetition of PTC

Repetition of PTC after a late recurrence yielded relief in 34 of 54 operations as follows: Abdennebi and Amber. 7 of 16: Esposito et al.. 7 of 18: Fraioli et al.. 8 of 12: Lichtor and Ylullan. 4 of 8: Lobato et al.. 8 of 11: and Peragut et al.. 8 of 9.1.24.26.46.47.6]

Major Sequelae

There have been two serious extratrigeminal sequelae. Dan (Per­sonal communication. 1988) has kindly described to me a man with persisting pain following severe Y] and Y 2 facial injuries without cranial abnormality. In the hope of achie\ing adequate numbness of the painful area the properly placed balloon was in­flated for 12 min. This was without incident until 8 h later when complete blindness came on in the ipsilateral eye. Optic atrophy continues at 2 years. The cause of this loss has not been deter­mined by ophthalmic. angiographic. and computed tomography (CT) studies.

The only death following this procedure. described by Spaziente et al.. occurred in a 62-year-old man with typical trigeminal neuralgia and a moderate asymptomatic hydrocephalus78 A larger than usual ( 12-gauge) needle "was not inserted beyond the foramen ovale" but clear CSF emerged when the stylet was re­moved. This stopped when the needle was retracted slightly and 0.7 ml of nonionic contrast medium injected to inflate the balloon to a characteristic pear shape for 6 min. yielding minimal sensory loss and lasting complete relief of pain. However. headache. drowsiness. and vomiting signaled a subarachnoid hemorrhage shown by CT the next day to fill the basal and sylvian cisterns and to increase the size of the ventricles. Cerebral angiography and studies of blood coagulopathy were normal. After a variety of vi­cissitudes, death finally ensued 7 months later. There seems to have been no diagnostic or technical error in either of these pa­tients. They illustrate the point that we in the United States are well advised to submit patients to any invasive procedure only when they have explicitly consented to it in writing with the understand­ing that any conceivable disaster may happen to them.

Arterial Injuries

The 14-gauge needle is a sizable spear to insert through the facial soft tissue. Lichtor and Mullan emphasize the importance of keep­ing it out of the intracranial cavity.46 Their assiduous analyses led in one patient with a fourth nerve paralysis of 3 months duration to the demonstration of a very small dural AYM. and in another pa­tient to an extracranial AYM. In a third case a maxillary artery fistula caused a persistent bruit . 'easily eliminated by endovascular wire coil occlusion." Revuelta et al. have added another case of arteriovenous fistula from the middle meningeal artery to the inter­nal jugular vein that closed spontaneously 19 days later as demon­strated by another angiogram66 They all suspect that in each case the large needle transfixed a small artery and vein. Lobato et al. were unable to place the balloon in Meckel's cave in one patient who shortly developed a low-flow carotid cavernous fistula that resolved spontaneously in 3 months.47 Mullan states that if inter­mittent active bleeding occurs during needle insertion it is desir­able to compress above the zygoma as well as below the maxilla.:i6 Lobato et al. had arterial bleeding through the cannula in four cases. into the external auditory canal in one. and from the nose in twO.46 They also state that the carotid or accessory meningeal artery traverses the foramen m'ale and that a tortuous carotid artery may pass directly over this foramen. They point out a number of other possible bony deficiencies in the region that may permit un­desirable arterial puncture. They recommend terminating the pro­cedure if arterial bleeding occurs. with which I heartily agree even when I am using a 20-gauge needle electrode. Meglio and Cionj report that bleeding not demanding cessation of the procedure oc­curred in three cases.:i2 The one of extracranial origin required both aural and nasal plugging: in another the puncture of the intra­cranial carotid was thought likely. The bleeding was venous in the third case. The outcome in these patients is not described. There is general agreement that a \enipuncture need not stop the operation.

Vasodepressor and Vasopressor Reflexes

Brown and Preul are so impressed with the degrees of bradycardia and hypotension seen during any phase of the procedure in most of the patients that they recommend not only continuous monitoring of arterial blood pressure and heart rate and availability of atropine for intravenous use at once. but also use of a responsive noninvasive temporary cardiac pacemaker throughout the procedure. This vasodepressor reflex has converted to a vasopressor response in four patients by blocking the third trigeminal division with lidocaine. Belber and Rak also advocate intravenous atropine. having used thoracic pressure resuscitation once in a patient with a 15-sec apnea Lichtor and Mullan describe several such episodes' 'with­out incident. Lobato et al.. on the other hand. noted "a steep increase in blood pressure in almost every patient upon inflation of the balloon. They counteract this with a predistension bolus of vasodilating sodium nitroprusside.

Other Minor Complications

Despite numerous instances of temporary unilateral masticator weakness, none has been permanent. However, Belber and Rak report one patient in whom a bilateral procedure performed in one sitting was followed by severe masticatory weakness of 8 days duration before the onset of recoveryY Frank and Fabrizi are the only ones to state that compression adequate to stop third-division pain is likely to cause excessive damage to the first and second divisions.27 A few extraocular palsies have all been temporary.

The dangers of manipulation in the mid-face of a 14-gauge needle and of inflating a balloon on the floor of the middle cranial fossa have been proven to be small and the chances of obtaining relief at the price of acceptable abnormal facial sensations are huge. The Mullan procedure is in my view clearly superior to intracisternal glycerol injections; his own results are especially impressive. The advantage of brief general anesthesia and minimal numbers with dysesthesia must be weighed against one fatality from hemorrhage and one case of unilateral blindness in patients treated without technical error. My colleagues and I are going to try to emulate Mullan's skill.

Procedure when PTC Fails

For those who experienced failure of percutaneous compressions who now wish the operation with the best chance of success, Mullan recommends partial root section of the portio major in the posterior fossa. A survey of the results beginning in 1929 with Waiter Dandy, who introduced that operation, finds them to be inferior to those of radiofrequency thermal rhizotomy, with respect to both relief of the paroxysmal pain and incidence of major extratrigeminal sequelae. Dandy himself gave up his partial division of the portio major because of his recurrences. Recently Klun described a nearly 50 percent recurrence rate following division of one-third of the portio major, the same procedure recommended by Brown and Mullan.

In Mullan's series of over 1000 trigeminal radiofrequency percutaneous rhizotomies and in at least six other series of about 1000 or more (Rhoton, 1985, and Thurel, 1987: Personal commu­nication), there have been no deaths. He have seen no permanent disability and no lasting deficit of any neural function other than that of the trigeminal nerve. He has not declined to operate on any patient, even the patient in renal or congestive heart failure. These seven groups have been able to provide pain relief in all but an average of I percent of their patients. I know of no operation in the posterior fossa that can equal this record.

To revert to the principal objective of this chapter. relation of the treatment to the cause of trigeminal neuralgia, this operation usually produces a temporary mild decrease or no change in objective tests of facial sensation following a 1-cc balloon inflation for only 1 min of neural compression, yet yields pain control superior or comparable to open vascular decompression in the juxtapontine region. I know of no evidence that neural indentation in that region is modified by such a moderate manipulation. Why these modest maneuvers stop the paroxysms of trigeminal neuralgia remains a complete puzzle. However, the fact that this ultra-low-risk maneuver has an excellent chance of stopping the pain seems to me to eliminate any justification for an open operation in the middle or posterior cranial fossa as the first invasive procedure in the treatment of trigeminal neuralgia.

Conclusions

The safest invasive manipulation is a percutaneous lesion made by (1) radiofrequency heat to produce hypalgesia. or (2) pure glycerol limited to a 10-min exposure if this produces anesthesia, or (3) inflation of a balloon in Meckel's cave for 1 min.

Inasmuch as an unusually conservative approach has an astonishing likelihood of success, it is not crucial at this point to resolve the question as to the cause or causes of the disorder. With respect to the selection of the first invasive treatment, it is immaterial whether the cause is abnormal myelin and/or axons of sensory trigeminal fibers, or significant extrinsic pressure against them, or both, or neither. If the initial conservative effort fails, there are several reasonable, more aggressive tactics to pursue. Clear-cut vascular compression of the trigeminal rootlets in the posterior fossa is a sufficiently uncommon cause of trigeminal neuralgia that an operation in the posterior fossa is not justified as the first invasive procedure in the treatment of this disorder.