Neurovascular compression and essential hypertension

Neuroradiology(1991) 33:2-8

euro--

radiology 9 Springer-Verlag1991

Originals Neurovascular compression and essential hypertension An angiographic study B. Kleineberg 1, H. Becker 1, and M. R. Gaab 2 Departments of 1 Neuroradiologyand 2 Neurosurgery,MedizinischeHochschuleHannover,Federal Republicof Germany Received: 29 March 1990

Summary. The pathogenesis of essential hypertension still remains unclear. Recently, it has been supposed, that an arterial compression of the left root entry zone (REZ) of the cranial nerves IX and X by looping arteries may play a pathogenetic role. In this report we verified this hypothesis retrospectively by vertebral angiographies in 99 hypertensive and 57 normotensive patients. The angiographic findings were compared with the results obtained from an anatomic study, in which the positions of 10 left vagus/glossopharyngeal nerves in the skull were radiographically determined in 10 cadavers. By using a pattern of R E Z topography developed from this information we obtained the following results: In 81% of the evaluable angiographies of hypertensive patients we found an artery in the left R E Z of cranial nerves IX and X. The normotensive patients showed an artery in the REZ only in 41.7% of cases. Our results support the hypothesis that essential hypertension may be combined with neurovascular compression of the left R E Z of cranial nerves IX/X. Key words: Vertebral angiography - Posterior fossa arteries - Root entry zone - Vagus and glossopharyngeal nerves - Essential hypertension - Neurovascular compression

For several years neurovascular compression has been widely accepted as a cause of trigeminal neuralgia and hemifacial spasm. In the majority of patients an ectatic and elongated artery has been found in association with the syndromes. In trigeminal neuralgia and hemifacial spasm, the neurovascular compression is mainly caused by the superior cerebellar artery and the anterior inferior cerebellar artery, respectively. Microsurgical vascular decompression has become an appropriate operative technique to relieve these syndromes [1-4]. Since 1973 Jannetta [5] has proposed a neurovascular compression of the left R E Z of cranial nerves IX and X as a possible cause of essential hypertension. Between 1975 and 1982 he operated on 36 hypertensive patients using

the microvascular decompression technique. Postoperatively the blood pressure returned to normal in 32 patients, in one case it improved significantly, and only 3 cases remained unchanged. Intraoperatively, the most frequently offending artery was the posterior inferior cerebellar artery (PICA), more rarely the vertebral artery (VA) and sometimes the PICA+ VA. Very seldomly other arteries caused compression for example the superior cerebellar artery or the basilar artery [5-7]. Jannetta confirmed his results in a nonhuman primate model. 5 animals were implanted with a pulsatile balloon in the root entry zone of the ninth and tenth cranial nerves on the left side. An increased blood pressure of 23 %-42% was observed after inflating the balloon [8]. In a post mortem study of Naraghi published in 1988 [9], all 24 patients with essential hypertension showed neurovascular compression of the R E Z of left cranial nerves IX and X. Neurovascular compression of the R E Z of left cranial nerves IX and X was not observed in any of 17 normotensive patients. Two of the normotensive patients had right sided compression. We report the retrospective evaluation of angiographies of hypertensive_patients. Further, we set out to provide anatomic and radiographic criteria to predict vascular compression preoperatively in patients with primary hypertension.

Materials and methods In order to localize the REZ of left cranial nerves IX and X in angiographies, 10 cadavers were examined, 5 males and 5 females between the ages of 40-74 years. Each cadaver had a normal brain, which was proved afterwards by neuropathological section. A left retromastoid craniectomy was performed on each cadaver. The left cerebellar hemisphere was carefully retracted using the operation microscope. After preparation of the jugular foramen the cranial nerves IX and X were identified. Three small arterial clips were

Fig. 1. Dorsal view of the posterior fossa. The cerebellum is retracted. Three clips are placed on the glossopharyngeal/vagus nerves. The important proximal clip is the very right one close to the brainstem. This clip marks the root entry zone (REZ). PO, pons; PICA, posterior inferior cerebellar artery; OL, olive; IX, glossopharyngeal nerve; X, vagus nerve; XI, accessorius nerve; JE jugular foramen; R, retractor Fig. 2. Radiographs in Towne (a) AP (b) and lateral (e) projection. Cranial nerves IX and X are marked by three clips. The proximal clip was set close to the brainstem and marks the root entry zone (arrows) placed on the nerves, one at the jugular foramen, one at the centre of the nerves and one very proximally close to the brain stem (Fig. 1). Most important was the proximal clip for this served to mark the R E Z of left cranial nerves IX and X. Finally, the dura and galea were replaced and sewn up in the original position. Radiographs were then made in lateral, Towne and anteroposterior (AP) projections (Fig. 2a-c), exactly appropriate to the clinical procedure during angiography. To identify angiographies of hypertensive patients, we retrospectively reviewed 806 patient reports ranging over 6years (1981-1987). 99vertebral angiographies were found of patients with definite essential hypertension, aged between 40 and 74 years (mean age 56.8 years), 59 from females and 40 from males. As a control group 57 angiographies of normotensive patients between 40-73 years (mean age 54.7years); 26 female and 31 male, were evaluated for comparison. Angiographies of three patients with definite renal hypertension were also analysed.

Results By superimposing all the clips in each projection, areas of 50-59 mm 2 were outlined. The proximal areas represent the R E Z of left cranial nerves IX and X (Fig. 3 a-c). In the Towne projection these clips are projected 10 mm to the left of the skull midline, and level with the inferior border of the internal acoustic meatus. The area of the clips in anteroposterior (AP) projection lies 10 mm to the left of the midline and 5 mm caudal to the inferior border of the internal acoustic meatus. In the lateral projection the clips appear 20 m m caudal to the petrous ridge and 15 mm dorsal to the clivus. F r o m this information overlay patterns were developed (Fig.4) for all three projections, Towne, A P and lateral. The angiographies could be easily evaluated using these patterns. In 64 of the hypertensive patients an artery was found in the R E Z . In 15 angiographies there was no artery in the

\

/

fFig. 3. All 10 clips together in each projection: Towne (a) AP (b) and lateral (c) projection. The proximal clips represent the root entry zone (arrows) R E Z , 20 angiographies could not be evaluated. For hypertensive patients, the evaluable angiographies comprised 81% with an artery in the R E Z and 19% without (Fig.S). T h e posterior inferior cerebellar artery (PICA) appeared most frequently in the R E Z (24 cases). In 17 angiographies the vertebral artery (VA) and in 9 both P I C A and VA crossed the R E Z . In 12 patients the anterior inferior cerebellar artery (AICA) was observed in the R E Z , in only 1 the basilar artery and in I the VA + A I C A (Fig. 6). F r o m the control group of 57 angiographies of normotensive patients, an artery in the R E Z was shown in 20. In 28 cases the R E Z was free of arteries , and 9 angiographies could not be evaluated. Again referring to the angiographies able to be evaluated, 42.7% showed an artery in the R E Z and 58.3% had no artery in the R E Z (Fig.5). In 10 cases the VA, in 6 cases the AICA, in only 3 cases the P I C A and in i case the VA + PICA crossed the R E Z (Fig.7). In 2 of the 3 renal hypertensive patients no artery lay in the R E Z , and in i the A I C A crossed the region.

Skull Midline

0

/J Petrous Ridge

0 ([~)

Inferior Border Internal Acoustic Meatus

AP Projection

Inferior Border Internal Acoustic Meatus

Towne Projection

Lateral Projection

Clivus

Fig.4. Overlay pattern including all three projections, Towne, AP and lateral

The differences between normotensive and hypertensive patients are statistically significant based on the )f-test for variation (Z2 = 20.64, P < 0.01). To obtain information on the influence of the different vessels in causing hypertension, logistic regression was applied [10]. This showed that hypertension occurs with a probability of 88.9% if the P I C A crosses the R E Z . Probability falls to 66.7% for the A I C A and to 65.5% for the VA. At 90% the probability of hypertension was highest if both the P I C A and AV appeared in the R E Z .

Discussion The ninth and tenth cranial nerves are positioned close to each other and leave the brainstem from the area retroolivaris a few millimeters caudally from the pons and about 3 mm dorsally from the olive. The two nerves run along a straight or slightly curved course towards the jugular foramen. The glossopharyngeal nerve leaves the skull

g Artery l u RBZ

~

O0

5G

4C

3G

212

IC

o HYPERTENS I V ES

N O R M O T E NISVES

Comparison of the root entry zones (REZ) of hypertensive and normotensive patients containing and not containing arteries. The proportion of evaluable angiographies is also shown Fig.5.

%

~

PICA

VA 50

~

PICA + VA

~AICA

40

/.......,.....: --7

30

. . . . . , ....... :+:+:.:.:.: ~::::::::::::::: ........:... :+:+:.:+:. .:.......:.. :+:.+:.:.: +:+:-:.:+ :.x+:+:.:

20

:!:i:!:!:i:i:i: ::::::::::::::::

i~i~!iiii~iiii!!

i0 +:.:.::.:.x

iiiiiii}iiiiii!iHYPERTENSIVES

NORMOTENSIVES

Fig.6. Distribution of different arteries in the root entry zone (REZ). PICA, posterior inferior cerebellar artery; VA, vertebral artery; AICA, anterior inferior cerebellar artery; BA, basilar artery

through the pars nervosa and the vagus nerve through the pars venosa of the jugular foramen [11]. The results of this study showed a fairly constant position of the ninth and tenth cranial nerves on the skull radiographs: The R E Z of the left cranial nerves IX and X, based on the addition of proximal clips, represent areas of only 50-59 mm 2, depending on different projections. The vertebral artery, located on average 8 mm distant from the front edge of the foramen magnum, enters the posterior fossa of the cranium. Frequently the left VA is of larger diameter than the right [12, 28]. On average 4 mm from the clivus, it courses to the caudal edge of the pons and merges with the VA on the other side to form the basilar artery [13]. In contrast to this usual course, the VA of the observed hypertensive patients, in lateral projec-

tion, described a curve of dorsal convexity (Fig.7 b). In Towne and AP projection even a loop is often formed towards the R E Z of the left cranial nerves IX and X (Fig.7 a). The BA emerges from the confluence of VAs on the anterior surface of the pons with its convexity against the clivus within the prepontine cistern; it extends as far as the interpeduncular cistern. The B A of hypertensive patients usually rises much higher than the dorsum sellae, sometimes even reaching the third ventricle. Where the VAs are of unequal diameter the B A usually shows an Sshaped course or a deviation from the midline contralateral to the larger VA [12, 13]. Because of the frequently larger left VA, the convexity of the caudal segment of the B A located at the level of the R E Z is usually right sided. So it is not surprising that the BA only once appeared in the left REZ. Among the arteries of the posterior fossa the PICA varies most in origin and distribution. In 18% of cases, the PICA arises below the foramen magnum, in 4% it originates at the level of the foramen magnum and in 57% above. In 21% its origin cannot be exactly defined, and in 20% no PICA exists. The anterior medullary segment of the PICA usually curves from the anterior surface of the medulla around the lower end of the olive. From here the lateral segment courses around the lateral medulla, often by forming a variable caudal loop [13-15]. The hypertensive patients in this study frequently showed a cranial loop of the anterior medullary segment towards the R E Z of the left cranial nerves IX and X (Fig.8a, b). Compared with the other arteries investigated, PICA appeared with the highest frequency in the R E Z of the left cranial nerves IX and X. In some cases even PICA and VA together crossed the R E Z (Fig. 9). 80-87% of the AICAs arise from BA. Within the pontocerebellar cistern close to the abducens nerve the A I C A courses laterally and caudally in the direction of the internal auditory canal, where it curves back medially and bi9furcates into a medial and a lateral branch [13]. Because of its mostly straight-line course as well as its relatively close anatomic relation to the cranial nerves IX and X the interpretation of AICA's appearance in the R E Z was problematic and requires further study. Numerous observations have shown that mechanical or electrical manipulations close to the brainstem, particularly in the region of its cardiovascular control centers, result in changes of blood circulation and especially in blood pressure. For many years the effects of intracranial bilateral glossopharyngeal nerve section have been known to vary from temporary elevation [16] to dramatic increase in blood pressure with fatal result [17]. Bilateral lesions of the nucleus tractus solitarii caused by syringomyelia or acute bulbar poliomyelitis associated with clinical hypertension have been described [18, 19]. Carefully directed lesions of the nucleus tractus solitarii confirmed these findings [20, 21]. Electrical stimulation of the brain stem, furthermore, led to an increase in blood pressure [22]. According to recent publications it is supposed that the most important region of the brain stem for regulation of blood pressure is the so-called Cl-adrenalin-cell group

Fig.7. Vertebral angiograms of a hypertensive patient in Towne (a) and lateral (b) projection. Frequent course of the vertebral artery when the artery is in the REZ. Note the S-shaped course of the BA with the convexity of its caudal segment to the right Fig.8. Towne (a) and lateral (b) view of the same hypertensive patient. The pattern of each projection is ovedayed. Both projections

show a typical loop of the PICA. The artery is definitely crossing the REZ Fig.9. Hypertensive patient with PICA and VA in the REZ of the left glossopharyngeal and vagus nerves in Towne (a) and lateral (b) view

(Fig. 10). Primary afferent fibres coming from baroreceptots, the aortic sinus and other cardiopulmonary receptors, via the ninth and tenth cranial nerves, terminate in the nucleus tractus solitarii. F r o m there, G A B A containing neurons project to the Cl-cell group, inhibiting it. The Cl-cell group sends projections into the intermediolateral column of the spinal cord, which innervate the adrenal medulla, and via the sympathic ganglia, heart and blood vessels [23-27]. Electrical or chemical stimulations of the Cl-cell group elicited a distinct elevation in blood pressure; lesions lead to decreased pressure [22, 27]. The Cl-cell group is situated at the ventrolateral medulla, exactly in the area of the R E Z of cranial nerves IX and X about i mm below the surface of the brain stem. An irritation of the Cl-cell group by pulsatile compression of an ectatic vessel would not be unlikely. In the clinical studies of Jannetta most of the hypertensive patients operated on by microvascular decompression of the R E Z of the left cranial nerves IX and X showed a significant decrease in blood pressure [8]. No changes in blood pressure occurred in patients operated on the right side. These results were confirmed surgically by Fein [28, 29] an anatomically by Naraghi [9]. In nonhuman models a small balloon pulsating synchronously with the heart rate was placed at the R E Z of cranial nerves IX and X of baboons, cats and dogs. The baboons with balloons placed on the left showed a significant elevation in blood pressure [8]. Pulsatile compression applied to cats' R E Z of the left cranial nerves IX and X induced an increased stroke volume and cardiac output [30]. U p o n compressing the right

f

X

x.

",

i i!l!

\\

Aortic \\ baroreceptors, \ \ afferent ~

Carotid sinus baroreceptors, afferent

x '

ii.:iliil / -x,, { ! I -'~t 1~ ~ ~,~ ~

// /

Intermediolateral column

Sympathetic pathway---> Heart, Vessels, Adrenal medulla Fig. 10. Sectional view of medulla oblongata at the level of the REZ of the glossopharyngeal nerve in the rat. C1, C1 arenalin cell groups; IO, inferior olivary nucleus; NTS, nucleus tractus solitarii; CST, corticospinal tract; IX, glossopharyngeal nerve; X, vagus nerve; XI, accessorius nerve; XII, hypoglossal nerve

R E Z of the cranial nerves IX and X in dogs only cardiac arrythmias without significant haemodynamic changes were observed [31]. It is well known that increasing age leads to elongation of the cerebral vessels. By forming a rostral loop the PICA, in particular, could approach the R E Z of the cranial nerves IX and X because of its long unfixed course. The effect of pulsatile compression to the R E Z could be an irritation of the Cl-cell group with resulting increase in blood pressure. Otherwise, interferences with vagal and glossopharyngeal afferents or efferents could lead to disorders in arterial baro- and chemoreceptor reflexes. It still remains unclear why only left-sided compression leads to elevation in blood pressure. There are some indications that the function of the vagus nerve is asymmetrical and that the left vagus nerve has the major control of the left heart. But there is confusion in the available literature [5, 31-39]. Our angiograhic findings coincide with the results of the Jannetta, Fein and Naraghi studies. We found an increased number of arterial vessels in the left R E Z of cranial nerves IX and X in primary hypertensive patients, which was of statistical significance. Even the comparative frequency of the offending vessels was the same, the most frequent being the PICA, followed by the VA. Interpretation of the A I C A was difficult, and probably this vessel alone was efficacious in some of the cases. The B A appeared only once in the REZ. This study supports the hypothesis of Jannetta that essential hypertension is caused by or associated with neurovascular compression of the R E Z of the left cranial nerves IX and X. The development of an overlay pattern to locate the R E Z for the left cranial nerves IX and X has provided a valuable diagnostic tool, and further prospective studies should follow. Moreover, the pattern can be used to predict neurovascular compression and so can be useful as an indicator of whether to operate by microvascular decompression technique. Beside these angiographic evaluations, magnetic resonance imaging should also be utilized. In a recently published case report vascular compression of the trigeminal nerve was demonstrated [40].

References 1. Jannetta PJ (1967) Structural mechanisms of trigeminal neuralgia: arterial compression of the trigeminal nerve at the pons in patients with trigeminal neuralgia. J Neurosurg 26:159-162 2. Jannetta PJ (1977) Treatment of trigeminal neuralgia by suboccipital and trm~stentorial cranial operations. Clin Neurosurg 24: 538-549 3. Haines SJ, Jannetta PJ, Zorub DS (1980)Microvaseular relations of the trigeminal nerve. J Neurosurg 52:381-386 4. Jannetta PJ, Abbasy M, Maroon JC, Ramos FM, Albin MS (1977) Etiology and definitive microsurgical treatment of hemifacial spasm. J Neurosurg 47:321-328 5. Jannetta PJ, Segal R (1985) Neurogenic hypertension: Etiology and surgical treatment. I. Observationes in 53 patients. Ann Surg 201:391-398

6. Jannetta PJ, Gendell HM (1978) Neurovascularcompression associated with essential hypertension. Neurosurgery 2:165 7. Jannetta PJ, Gendell HM (1979) Clinical observations on etiology of essential hypertension. Surg Forum 30:431-432 8. Jannetta PJ, Segal R, Wolfson SK, Dujovny M, Semba A, Cook EE (1985) Neurogenic hypertension: Etiology and surgical treatment. II. Observations in an nonhuman primate model. Ann Surg 202:253-261 9. Naraghi R, Gaab MR, Walter GF (1988) Neurovascular compression as an etiology of essential hypertension. A microanatomical study. In: Advances in neurosurgery. 39th annual meeting of the german neurosurgical society, Cologne 10. Cox DR (1970) The analysis of binary data. Methuen, London 11. Lang J (1982) Uber Ban, L~inge und Gefal3beziehungen intrazisternaler Hirnnerven. Zbl Neurochir 43:244247 12. Kazui S, Kuriyama Y, Naritomi H, Sawada T, Ogawa M, Maruyama M (1989) Estimation of vertebral arterial asymmetry by computed tomography. Neuroradiology 31:237-239 13. Krayenbfihl H, Yasargil MG (1979) Zerebrale Angiographie flir Klinik und Praxis, 3. Aufl. Thieme, Stuttgart 14. Margolis MT, Newton ThH (1972) Borderland of the normal and abnormal posterior inferior cerebellar artery. Acta Radiol Diagn 13:163-172 15. Margolis MT, Newton ThH (1974) The posterior inferior cerebellar artery. In: Newton ThH, Potts DJ (eds) Radiology of the skull and brain, Book 2, 1st edn. Mosby, St. Louis 16. Whisler WW, Voris HC (1965) Effect of bilateral glossopharyngeal nerve section on blood pressure. A case report. J Neurosurg 23:79~81 17. Wycis H (1945) Bilateral intracranial section of the glossopharnygeal nerve. Report of a case. Arch Neurol Psychiatry 54: 344347 18. Magnus O, Koster M, Van Der Drift JHA (1977) Cerebral mechanisms and neurogenic hypertension in man: With special reference to baroreceptor control. In: De Jong W, Provost AP, Shapiro AP (eds) Hypertension and brain mechanisms. Progress in brain research, vo147. Elsevier, Amsterdam, p 199 19. Baker AB, Matzke HA, Brown JR (1950) Poliomyelitis. III. Bulbarpoliomyelitis: a study of medullary function. Arch Neurol Psychiatry 63:257 20. Reis D J, Doba N, Nathan MA (1976) Neurogenic arterial hypertension produced by brainstem lesion. In: Onesti G, Fernandes M, Kim KE (eds) Regulation of blood pressure by the central nervous system. Grune & Stratton, New York San Francisco London, pp 35-51 21. Snyder DW, Doba N, Reis DJ (1978) Regional distribution of blood flow during arterial hypertension produced by lesions of the nucleus tractus solitarii in rats. Circ Res 42:87-91 22. Peiss CN (1958) Cardiovascular response to electrical stimulation of the brain stem. J Physio1141:500-509 23. Armstrong DM, Ross CA, Pickel VM, Job TH, Reis DJ (1982) Distribution of dopamine-, noradrenalin-, and adrenalincontaining cell bodies in the rat medulla oblongata: Demonstrated by the immunocytochemical localisation of catecholamine biosynthetic enzymes. J Comp Neuro1212:173-187 24. Ross CA, Ruggiero DA, Joh TH, Park DH, Reis DJ (1983) Adrenalin synthesizing neurons in the rostral medulla: a possible role in tonic vasomotor control. Brain Res 273:356-361

25. Reis D J, Ross CA, Ruggiero DA (1984) Role of adrenaline neurons of ventrolateral medulla (the C1 group) in the tonic and phasic control of arterial pressure. Clin Exper Hyper Theory Praxis 6:221-241 26. Benarroch EE, Granata AR, Ruggiero DA, Park DH, Reis DJ (1986) Neurons of C1 area mediate cardiovascular responses initiated from ventral medulla surface. Am J Physio1250: R932R945 27. Ciriello J, Caverson MM, Polosa C (1986) Function of the ventrolateral medulla in the control of the circulation. Brain Res Rev 11:359-391 28. Fein JM (1981) Hypertension and central nervous system. Clin Neurosurg 29:666-721 29. Fein JM, Frishman W (1980) Neurogenic hypertension related to vascular compression of the lateral medulla. Neurosurgery 6: 615~522 30. Segal R, Gendell HM, Canfield D, Dujovny M, Jannetta PJ (1982) Hemodynamic changes induced by pulsatile compression of the ventrolateral medulla. Angiology 33:161-172 31. Segal R, Jannetta PJ, Wolfson SK, Dujovny M, Cook EE (1982) Implanted pulsatile balloon device for simulation of neurovascular compression syndroms in animals. J Neurosurg 57:646-650 32. Dawes J, Widdicombe JG (1953) The afferent pathway of the Bezold reflex: The left vagal branches in dogs. Br J Pharmacol 8: 395-398 33. Faden AI, Jacobs TP, Woods M, Tyner CF (1978) Zona intermedia pressor sites in the cat spinal cord: right-left asymmetry. Exper Neuro161:571-581 34. Randall WC (1984) Selective automatic innervation of the heart. In: Randall WC (ed) Nervous control of cardiovascular function. Oxford University Press, New York Oxford, pp 46-67 35. Gunn GG, Sevelius G, Puiggari MJ, Myers FK (1968) Vagal cardiomotor mechanisms in the hindbrain of dog and cat. Am J Physio1214:258-262 36. Henry JL, Calaresu FR (1974) Exitatory and inhibitory inputs from medulla nuclei projecting to spinal cardioacceleratory neurons in the cat. Exp Brain Res 20:485-504 37. Henry JL, Calaresu FR (1974) Pathways from medulla nuclei to spinal cardioacceleratory neurons in cat. Exp Brain Res 20: 506514 38. Stuesse SL (1982) Origins of vagal preganglionic fibres: a retrograde transport study. Brain Res 236:15-25 39. Szulczyk P, Tafil M (1980) Influence of the input from left and right arterial barorecteptors on left inferior cardiac nerve activity in cats. J Auton Nerv Syst 2:355-364 40. Wong BY, Steinberg GK, Rosen L (1989) Magnetic resonance imaging of vascular compression in trigeminal neuralgia. J Neurosurg 70:132-134

Dr. B. Kleineberg Abteilung Neuroradiologie Medizinische Hochschule Hannover Konstanty-Gntschow-Strage 8 W-3000 Hannover 61 Federal Republic of Germany