Histological Study of the Normal Vertebral Artery —Etiology of Dissecting Aneurysms—

Neurol Med Chir (Tokyo) 44, 629¿636, 2004

Taku SATO,TatsuyaSASAKI,KyouichiSUZUKI,MasatoMATSUMOTO, Namio KODAMA,andKouichiHIRAIWA*

Departments of Neurosurgery and *Legal Medicine, Fukushima Medical University, Fukushima

Abstract Histological evaluation of dissecting aneurysms of the cerebral arteries has suggested that defects in the internal elastic lamina (IEL) induce dissection of the arterial wall. Dissecting aneurysms occur most frequently in the vertebral artery (VA). The present study examined sections of the normal VA to elucidate the mechanisms of arterial dissection. Bilateral VAs (20 vessels) were obtained from 10 patients who died of causes other than intracranial lesions. The VAs were detached from the VA union to the site 10 mm proximal from the point penetrating the dura mater. The VAs were cut at 5-mm intervals, and each segment was observed using modified Masson's trichrome staining for elastic fibers. The thickness of the media and adventitia significantly thinned after the origin of the posterior inferior cerebellar artery (PICA). IEL defects were observed at 35 sites in 11 vessels from six subjects. There was a high incidence of IEL defects in the extradural portion and near the origin of the PICA, areas frequently involved in arterial dissection. There was a high incidence of intimal thinning at areas of IEL defects (19 of 35), and thinning was particularly marked distal to the origin of the PICA. In the absence of intimal thickening, the vascular strength at the site of IEL defects may be reduced, which would promote the occurrence and progression of arterial dissection.

Key words: histology, vertebral artery, dissecting aneurysm, elastic lamina

Introduction The etiology and pathology of dissecting aneurysms of the cerebral arteries remain largely Advances in imaging techniques have led to an unknown. The internal elastic lamina (IEL) is the increase in the detection of intracranial dissecting strongest structure in the arterial wall and IEL aneurysms. Dissecting aneurysms of the carotid defects may be a primary factor in intimal rup- artery system often result in cerebral infarction due ture.15,16) Histological evaluation of dissecting to arterial stenosis or obstruction. Dissection of aneurysms of the cerebral arteries has suggested the vertebral artery (VA) system often leads to that defects in the IEL or media, or degeneration of subarachnoid hemorrhage with high rebleeding the media induce intimal changes, and rupture of rate and mortality.1,3,14,27) Slight injury due to neck the intima results in dissection of the arterial exercise or chiropractic manipulation9,22) may cause wall.1,4,6,7,15,17–21) Other dissections were attributed to VA dissection. Arteriosclerosis, fibromuscular rupture of the vasa vasorum or a newly formed ves- dysplasia,2,19) and degeneration of the media may be sel in the vascular wall associated with degeneration related to arterial dissection. Extracranial dissection of the media.5,28) The intima and IEL at the opening of the VA occurs most frequently at the C1–2 levels of a dissection are washed away by blood, so of the spine,22) whereas intracranial dissections are preventing histological investigation. Dissecting often located near the origin of the posterior inferior aneurysms occur most frequently in the VA,25) but cerebellar artery (PICA).10,18,26) Minimal neck injury few histological investigations of the normal VA has been suggested as a cause of dissecting have investigated the underlying etiologic mechan- aneurysms at the C1–2 levels,22) but the preferential isms. In particular, the thickness of the wall of the occurrence near the origin of the PICA remains to cervical and intracranial VA near the dura mater12,24) be elucidated. and the presence of IEL defects are not well charac- terized. Received April 19, 2004; Accepted August 9, 2004 The present histological evaluation of the vascular

629 630 T. Sato et al.

Table 1 Summary of cases

Case Age Sex Cause of death No. (yrs)

1 0.7 M sudden infant death syndrome 2 27 M alcoholism 3 28 F pulmonary infarction 4 35 M bleeding due to injury of the jugular vein 5 51 M rupture of the thoracic aorta 6 52 M heart failure 7 64 F bleeding due to liver laceration 8 67 M heart failure 967Mhypothermia 10 92 F heart failure

structure of the normal VA measured the thickness of the three-layer structure in the apparently normal Fig. 1 Schematic drawing of the vertebral artery VA and studied the distribution of the IEL and the specimens. The point 10 mm toward the condition of the intima at IEL defects to clarify the heart from the site where the vessel pene- etiology and pathology of dissecting aneurysms of trates the dura mater (dura) was defined as the cerebral artery. A, the point 5 mm distal to A as B, and the site of penetration of the dura mater as C. Materials and Methods Points D–J are at 5-mm intervals from point C in the intradural portion. Serial sections 4 mminthickness(n＝ 20) were prepared at Bilateral VAs (20 vessels) were obtained from 10 points A–J. PICA: posterior inferior patients, aged 8 months to 92 years, who died of cerebellar artery. causes other than intracranial lesions and under- went legal autopsy at our hospital (Table 1). The bilateral VAs were detached intradurally from the ventral, lateral, dorsal, and medial aspects of the union (origin of the basilar artery) retrogradely to vascular cross-sections were labeled a, b, c, and d to where the VA penetrated the dura mater, and were facilitate orientation of each specimen. Since the resected extradurally en bloc from where the VA presence of the PICA, which usually branches first penetrated the dura mater to 10 mm toward the from the intracranial VA, might affect the measure- heart. ment results, the presence and the position of the Detached VAs were cut at 5-mm intervals, and branching of the PICA relative to point C were each segment was fixed in formalin and embedded recorded. in paraffin. The point 10 mm toward the heart from The thicknesses of the intima (from the surface of the site where the VA penetrated the dura mater was the endothelial cells to the external aspect of the defined as A, the point 5 mm distal to A as B, and the IEL), the media (thickness of smooth muscle fibers), site of penetration of the dura mater as C. Points D–J and the adventitia (thickness of elastic fibers and were assigned at 5-mm intervals from point C collagen fibers) were measured in aspects a, b, c, and (Fig. 1). Twenty 4 mm-thick sections were prepared d at points A to J in the 20 vessels. The mean value of at points A–J and stained for elastic fibers using the four aspects was calculated for each layer at modified Masson's trichrome staining for examina- points A–J. Since the length of the VA in the infant tion under a light microscope. Due to variations in (Case 1) differed greatly from the adult cases, the length of the VAs among the 10 subjects, six measurements were made only at points where there blocks were prepared in two vessels, seven blocks in was clear correspondence with adult VAs (e.g. point one vessel, eight blocks in 11 vessels, nine blocks in C). three vessels, and 10 blocks in three vessels. The IEL defects with lengths of 25 mmorlonger,which ventral aspect of the cross-section of the artery was were detectable by light microscopy, were regarded marked with India ink at harvest. The surface of the as defects (Fig. 2), and the frequency was compared basilar artery which came into contact with the at points A–J and aspects a–d in the 10 subjects. The brainstem was regarded as the dorsal aspect. The thickness of the intima at each IEL defect was deter-

Neurol Med Chir (Tokyo) 44, December, 2004 Histology of the Normal VA 631

Fig. 2 Photomicrograph showing a defect (aster- isk) in the internal elastic lamina. Defects 25 mm or longer that were detectable by light microscopy were considered defects. Mo- dified Masson's trichrome stain for elastic fibers, ×25. mined as the mean longitudinal diameter by dividing the area of the defect by its transverse diameter. The intimal area was measured using area-calculat- ing software (Scion Image; Scion Corp., Frederick, Md., U.S.A.). The intimal area was used because the thickness of the intima sometimes varied widely between the start and the end of the IEL defect, so the mean longitudinal diameter yielded a more accurate representation. The degree of thickening of Fig. 3 Thicknesses of the intima (△), media (), the intima at each IEL defect was calculated by and adventitia () of the 20 vertebral dividing the thickness of the intima at the defect by arteries. Values are mean ± standard the thickness of the normal intima, which was error. PICA: posterior inferior cerebellar determined in advance as the reference value. artery. Thevaluesareshownasmean± standard error. Analysis of variance (Scheffe's post-hoc test) was performed at the p º 0.05 level of significance for was greatest (253 ± 10 mm) at point B and decreased comparisons between sites. most sharply after point C, reducing to 192 ± 10 mm at point D. More distally, the thickness decreased Results more gradually. The media was significantly thinner at point G than point B. The thickness of the adven- The PICA originated at the VA between points F and titia showed no significant differences at aspects G in 17 of the 18 specimens from the nine adults. a–d. The thickness of the adventitia was greatest The mean distance from point C to the origin of the (247 ± 52 mm) at point C, decreased after point C to PICA was 15.9 ± 2.0 mm (range 0–33 mm). 128 ± 11 mm at point D, and thinned more gradually The thickness of the intima, media, and adventitia further distally. The adventitia was significantly of the 20 VAs is shown in Fig. 3. There was no sig- thinner at points G–I than point A or B, at points E–J nificant difference in the thickness of the intima at than point C, and at points F–I than point D. aspects a–d. The mean intimal thickness of the four IEL defects were observed at 35 sites in 11 vessels aspects was 60 mm or less at, or proximal to, point C, from six subjects aged 51 years or older (Table 2), 85 mm or more at points D–F, and 65 mmorlessat but not in the four subjects aged 35 years or younger. point G. The thickness of the intima did not differ Case 10, a 92-year- old woman, had IEL defects at 14 significantly at any of the points of measurement. sites (Fig. 4 upper). The distribution of the defects The thickness of the media was not significantly showed 11 at point A or B, and 17 at point E, F, or G different at aspects a–d. The thickness of the media (Fig. 4 lower). Defects were located at aspects a and

Neurol Med Chir (Tokyo) 44, December, 2004 632 T. Sato et al.

Table 2 Degree of intimal thickening at the defect in the internal elastic lamina

Case Length of defect Site Intimal thickness No. (mm)

5 A 86 1.77 A 432 3.71 B 1008 2.88 B 864 1.22 B 860 1.15 B 140 0.99 F 886 0.42 G 234 0.08 G 878 0.59 6 E 26 0.60 G520.40 7 C 650 0.71 E 231 9.28 G 286 0.60 G 246 0.69 8 A 300 3.91 E601.99 F 2100 2.05 9 E 169 0.43 G641.05 H261.07 10 A 450 0.31 A 1588 0.86 A 600 0.76 A 185 0.83 D 618 0.58 F 133 1.10 F 1330 1.34 F 106 0.87 G820.98 G800.42 H780.42 H800.53 J 1025 1.51 J 213 0.35

b in 10 sections, c in four sections, and d in 11 sections. There was no significant difference in the regional distribution of these defects. The degree of intimal thickening was less than 0.9, or the intima was thinner than normal (Fig. 5 upper left), at 19 of the 35 IEL defects; 0.9–1.1, or nearly the same as in normal vessels, at five defects; and much greater than 1.1, or marked (Fig. 5 lower), at 11 defects (Table 2). The intima was thickened at five of Fig. 4 Distribution of defects in the internal elastic the 10 extracranial sites (points A and B), and was lamina (lower)shownasdotsat35sitesin thicker than the intima of normal vessels at three of 11 vessels from six subjects aged 51 years or 16 intracranial sites (points F–H). older (upper).

Neurol Med Chir (Tokyo) 44, December, 2004 Histology of the Normal VA 633

thinner intracranially than extradurally in the study population, but the difference was not remarkable and few age-related changes were noted. The adventitia was thicker intracranially, but measurements were obtained at only one unidentified site of the intracranial VA. The present study examined the VA from 10 mm proximal to the site of penetration of the dura mater to the union of the intracranial VAs by serial section- ing at 5-mm intervals. The intima thickened after penetration of the dura mater and thinned after the branching of the PICA. The media thinned after penetration of the dura mater and was particularly thin distal to the branching of the PICA. We also found that the adventitia was thicker than the media just after penetration of the dura mater and thinned after the branching of the PICA. We suggest that the high incidence of rupture and bleeding near the branching of the PICA, regardless of the position of the opening of the dissection lumen, may be related to the observed medial and adventitial thinning at this site. Pathological study of a large number of autopsy cases with VA dissecting aneurysms suggested that vascular dissection is attributable to IEL rupture.15) Elastin, a structural protein that provides strength to the vascular wall, is scarcely present in the media or adventitia, but is concentrated in the IEL, which is considered to be the strongest structure of the vascular wall.23) Ligation of major cerebral arteries obtained at autopsy at both edges and exposure to con- tinuous air flow pressure found that the artery with Fig. 5 Photomicrographs showing defects (aster- the intima and IEL tolerated up to 600 mmHg.8) A isks) in the internal elastic lamina. The degree of intimal thickening was less than postmortem study of human radial arteries showed 0.9 (upper left) and much greater than 1.1 that the IEL was a single- or multiple-layer structure (lower). Modified Masson's trichrome in young individuals but was stretched, ruptured, or stain for elastic fibers, ×10 (upper right) absent in some areas in older persons.11) and ×100 (upper left and lower). The present study observed IEL defects at 35 sites in 11 vessels from six subjects, all of whom were aged 51 years or older. Although our study popula- Discussion tion was small and age varied widely, the results suggest that the incidence of IEL defects increases Histological examination of the VA 1 cm proximal with age. Defects were located more frequently and distal to the site of penetration of the dura mater before penetration of the dura mater, and near the found that the adventitia and media thinned to branching of the PICA at 10–20 mm distal to the site approximately half the thickness after penetration of penetration of the dura mater. The high incidence butthatthethicknessoftheintimaandIEL of defects in these areas coincides with clinical remained unchanged.24) Histological study of nine observations that the VA at the C1–2 levels of the sites in the normal cervical VA and one site in the spine appears to be particularly prone to dissection intracranial VA showed that the mean thickness of and obstructive circulatory impairment following the intima at the 10 sites increased with age.12) Since mild trauma because of the mobility of the structures the thickness of the media was relatively unchanged, in that region,22) and that dissecting VA aneurysms the ratio of intimal:medial thickness was examined frequently occur and often bleed near the branching and found to be 1:30–1:50 in young persons and 1:2 of the PICA.10,18,26) WesuggestthatIELdefects in individuals by the age of 60 years. The media was before penetration of the dura mater and near the

Neurol Med Chir (Tokyo) 44, December, 2004 634 T. Sato et al. branching of the PICA may be related to the develop- Acknowledgments ment of dissection at these sites. Experimental IEL defects in rabbit femoral artery We are grateful to Mr. Takeshi Itakura for technical were repaired by local intimal thickening.13) assistance. However, intimal thickening due to atherosclerosis canalsoleadtoIELdefects.Infact,weobserved References marked intimal thickening at 11 of the 35 sites with IEL defects and usually thin intima at 19 sites. Histo- 1) Adams HP Jr, Aschenbrener CA, Kassell NF, logical examination disclosed intimal thickening Ansbacher L, Cornell SH: Intracranial hemorrhage and a few atherosclerotic changes such as macro- produced by spontaneous dissecting intracranial phage infiltration at most defective IEL sites. In- aneurysm. Arch Neurol 39: 773–776, 1982 timal thickening was observed at extracranial sites 2) Anson J, Crowell RM: Cervicocranial arterial dissec- (points A and B) but not at intracranial sites (points tion. Neurosurgery 29: 89–96, 1991 F–H) despite the high incidence of IEL defects. This 3) Aoki N, Sakai T: Rebleeding from intracranial dis- may be attributable to differences in the blood flow secting aneurysm in the vertebral artery. Stroke 21: 1628–1631, 1990 in the VA. The VA leaves the intervertebral foramen 4) Berger MS, Wilson CB: Intracranial dissecting of C-1, runs a transverse course, then curves nearly aneurysms of the posterior circulation. Report of six perpendicularly, and finally follows an anterior cases and review of the literature. J Neurosurg 61: course. Since point A was located immediately after 882–894, 1984 the perpendicular curve, the observed intimal thick- 5) Endo S, Nishijima M, Nomura H, Takaku A, Okada ening may be attributable to biological reactions in E: A pathological study of intracranial posterior response to turbulence and eddy currents in the circulation dissecting aneurysms with subarachnoid blood flow. On the other hand, intimal thinning at hemorrhage: Report of three autopsied cases and sites of IEL defects, such as points F–H, is indicative review of the literature. Neurosurgery 33: 732–738, of weakening of the vascular wall. An event such as 1993 intimal detachment may readily lead to arterial dis- 6) Farrell MA, Gilbert JJ, Kaufmann JC: Fatal intracranial artery dissection: Clinical pathological section, rupture, and hemorrhage. correlation. J Neurol Neurosurg Psychiatry 48: Thethicknessoftheintima,media,andadventitia 126–129, 1985 of normal VAs showed characteristic regional 7) Friedman AH, Drake CG: Subarachnoid hemorrhage changes from the extradural portion to the union of from intracranial dissecting aneurysm. J Neurosurg the bilateral VAs, and the media and adventitia 60: 325–334, 1984 significantly thinned after the branching of the 8) Glynn LE: Medial defects in the circle of Willis and PICA. Considering that rupture and bleeding occur their relation to aneurysm formation. J Pathol Bac- frequently at this site, thinning of the media and teriol 51: 213–222, 1940 adventitia may be involved in the vulnerability of 9) Haldeman S, Kohlbeck FJ, McGregor M: Risk factors this area to rupture following vascular dissection. A and precipitating neck movements causing ver- high incidence of IEL defects was found in the tebrobasilar artery dissection after cervical trauma extradural portion and near the branching of the and spinal manipulation. Spine 24: 785–794, 1999 10) Hasue M, Fukuda A, Nakajima S, Fukushima C, PICA in normal VAs, areas frequently involved in Haraoka J, Ito H: [Treatment of dissecting vertebral arterial dissection. There was a high incidence of aneurysms]. Surgery for Cerebral Stroke 21: 361–367, intimal thinning at areas of IEL defects (19 of 35), 1993 (Jpn) and thinning was particularly marked distal to the 11) Iwao Z: [Pathological study of arteriosclerosis of the branching of the PICA. The vascular strength at the radial artery]. Nagasaki Igakkai Zasshi 36: 379–393, site of IEL defects may be reduced in the absence of 1961 (Jpn) intimal thickening, and this may be a factor in the 12) Johnson CP, Lawler W, Burns J: Use of histomorpho- occurrence and progression of arterial dissection. metry in the assessment of fatal vertebral artery IEL defects with intimal thickening were more dissection. JClinPathol46: 1000–1003, 1993 frequent in the extracranial segment of the VA. 13) Mitchell GM, MaCann JJ, Rogers IW, Hickey MJ, Morrison WA, O'Brien BMC: A morphological study Since the extracranial VA curves nearly perpendicu- of the long-term repair process in experimentally larly twice immediately before penetration of the stretched but unruptured arteries and veins. Br J dura mater, intimal thickening at that site is a result Plast Surg 49: 30–34, 2001 of biological reactions to marked blood flow turbu- 14) Mizutani T, Aruga T, Kirino T, Miki Y, Saito I, lence and eddy currents. Tsuchida T: Recurrent subarachnoid hemorrhage from untreated ruptured vertebrobasilar dissecting aneurysm. Neurosurgery 36: 905–913, 1995

Neurol Med Chir (Tokyo) 44, December, 2004 Histology of the Normal VA 635

15) Mizutani T, Kojima H, Asamoto S, Miki Y: Pathologi- 23) Stehbens WE: Pathology of the Cerebral Blood cal mechanism and three-dimensional structure of Vessels. St Louis, CV Mosby, 1972, pp 60–97 cerebral dissecting aneurysms. JNeurosurg94: 24) Wilkinson IMS: The vertebral artery. Extracranial 712–717, 2001 and intracranial structure. Arch Neurol 27: 392–396, 16) Mizutani T, Miki Y, Kojima H, Suzuki H: Proposed 1972 classification of nonatherosclerotic cerebral fusiform 25) Yamaura A: Non traumatic intracranial arterial and dissecting aneurysms. Neurosurgery 45: 253–260, dissection. Natural history, diagnosis, and treatment. 1999 Contemp Neurosurg 16: 1–6, 1994 17) Nass R, Hays A, Chutorian A: Intracranial dissecting 26) Yamaura A, Watanabe Y, Saeki N: Dissecting aneurysms in childhood. Stroke 13: 204–207, 1982 aneurysms of the intracranial vertebral artery. J 18) Ono J, Yamaura A, Kubota M, Hirai S, Miyata A: Neurosurg 72: 183–187, 1990 [Management of 42 patients with ruptured dissecting 27) Yasui T, Kishi H, Komiyama M, Iwai Y, Yamanaka K, aneurysms in vertebrobasilar system. Significance of Nishikawa M, Nakajima H, Morikawa T: [Rerupture rebleeding]. Surgery for Cerebral Stroke 24: 51–56, mechanism of ruptured intracranial dissecting 1996 (Jpn) aneurysm in the vertebral artery following proximal 19) Pilz P, Hartjes HJ: Fibromuscular dysplasia and occlusion]. No Shinkei Geka 28: 345–349, 2000 (Jpn, multiple dissecting aneurysms of intracranial arte- with Eng abstract) ries. Stroke 7: 393–398, 1976 28) Yonas H, Agamanolis D, Takaoka Y, White RJ: 20) Sasaki O, Ogawa H, Koike T, Koizumi T, Tanaka R: A Dissecting intracranial aneurysms. Surg Neurol 8: clinicopathological study of dissecting aneurysms of 407–415, 1977 the intracranial vertebral artery. JNeurosurg75: 874–882, 1991 21) Scott GE, Neubuerger KT, Denst J: Dissecting aneurysm of intracranial arteries. Neurology 10: Address reprint requests to:T.Sato,M.D.,Departmentof 22–27, 1960 Neurosurgery, Fukushima Medical University, 1 22) Shimizu J, Nakagawa Y, Fuji Y, Nakase H, Mannen Hikarigaoka, Fukushima 960–1295, Japan. T: [Wallenberg's syndrome due to vertebral artery e-mail:tak-s＠fmu.ac.jp dissection following minimal neck injury. Report of two cases]. Rinsho Shinkeigaku 32: 430–435, 1992 Commentary on this paper appears on the next page. (Jpn, with Eng abstract)

Neurol Med Chir (Tokyo) 44, December, 2004 636 T. Sato et al.

Commentary strengthofthearterialwallandapossiblesitethat will promote the occurrence and progression of a Sato et al. present an interesting histological evalua- vertebral dissection. All the subjects with these altera- tion of the vertebral artery wall in non-dissected tions were aged over 50 years old, suggesting that the subjects as an attempt to explain the most frequent IEL defects increase with age. The IEL defect with sitesofdissectioninthisartery.Theextradural intimal thickening observed in the extracranial portion of the vertebral artery and the portion of the portion of the VA can be a reparative mechanism artery located near the origin of the PICA were the reactive to the flow turbulence in the curves of this principal sites of defects in the internal elastic lamina artery at that level. This paper is an attractive effort to (IEL) observed in 55% of the vessels studied. These form an eventual explanation for the prevalence of sites are involved frequently in vertebral artery dissec- some sites among others in VA dissections. tions, as pointed out by the authors. In 54% of the Jorge MURA, M.D., David ROJAS-ZALAZAR,M.D., cases the defect zones in the IEL mentioned above and Evandro de OLIVEIRA,M.D. appear concurrently with thinning of the intimal Instituto de Ci âencias Neurol áogicas layer. These two alterations can determine reduced S ãao Paulo, Brazil

Neurol Med Chir (Tokyo) 44, December, 2004