J7ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:17-21

PAPERS

Mechanisms of cerebral artery thrombosis: a

histopathological analysis on eight necropsy cases

Jun Ogata, Junichi Masuda, Chikao Yutani, Takenori Yamaguchi

AbstractThe mechanisms of thrombus formationin atherosclerotic cerebral arteries are

still controversial, although intraplaquehaemorrhage and rupture of the athero-matous plaques have been proposed. Ahistopathological analysis of segments ofthe thrombosed large intracranial arter-ies was carried out on eight patients whodied within 28 days after brain infarc-tion. The study revealed occlusivethrombi in six and mural thrombi in two,developing mostly at the site of greateststenosis or just distal to it. The histologi-cal characteristics of the thrombosedarteries were plaque rupture in three,intramural haemorrhage in one, ulcera-tion in one, and thrombosis in theabsence of plaque rupture or intramuralhaemorrhage in three. Occlusive embolidistal to the site of cerebral arterythrombosis (intracranial artery-to-arterythromboembolism) were observed intwo. The following conclusions were

drawn from the study: 1) plaque ruptureor intraplaque haemorrhage is not a sine

qua non for cerebral artery thrombosis;2) occlusions occur at the site of greatestluminal compromise or just distal to it,and 3) non-occlusive mural thrombosiscan occur in the absence of plaque rup-ture and eventually lead to local occlu-sion or distal embolisation.

Research Institute,NationalCardiovascularCenter, Osaka, JapanJ OgataJ MasudaDepartment ofPathologyC YutaniCerebrovascularDivision, DepartmentofMedicineT YamaguchiCorrespondence to:Dr Ogata, NationalCardiovascular Center,Research Institute, 5-7-1,Fujishirodai, Suita, Osaka,JapanReceived 9 October 1992and in revised form11 January 1993.Accepted 19 January 1993

(7 Neurol Neurosurg Psychiatry 1994;57: 17-2 1)

Because of the favourable outcome inpatients with large intracranial artery throm-bosis, there are few histopathological studieson this subject. Paterson observed that inti-mal haemorrhage due to rupture of capillariesderived from the main arterial lumens -wasintimately concerned with the mechanism ofcerebral artery thrombosis in six cases.'

Constantinides found that thrombi in largebranches of the circle of Willis in 10 cases

were attached to breaks in atherosclerotic or

fibrosed arterial wall.2 Sadoshima et alreported that intramural haemorrhage fromsmall intramural blood vessels was the majorcause of thrombosis in 39 thrombosed arterialsegments of branches in the circle of Willis.3

There has been no other histopathologicalstudy on the mechanisms of large intracranialartery thrombosis.4

Conversely, abundant histopathologicalstudies have been performed on coronaryartery thrombosis in myocardial infarction,and have documented a rupture of atheromaas a major cause leading to thrombus forma-tion.5-'2 Though break,28 rupture,'-'0 fissur-ing5 811 12 and cracking8 of atheroscleroticplaque of the cerebral or coronary arteries areterms that have appeared in previous reportsdescribing the disruption of the luminal sur-face of the plaque, we have used the term"rupture of the plaque" in this report.The necropsied patients died within 28

days after brain infarction due to large intra-cranial artery thrombosis and we were able toperform detailed histopathological analysis onthe mechanisms leading to the thrombus for-mation.

Subjects and methodsAmong 771 adult brains necropsied at theNational Cardiovascular Center between1979 and 1991, 114 brains had infarcts whichhad developed within 28 days before death.Of these cases, the intra-arterial thrombi inthe circle of Willis and its major brancheswere histopathologically confirmed in only sixcases, after thromboemboli originating fromthe heart or the carotid arteries and thrombiassociated with aneurysms of the cerebralarteries were excluded. Two additionalnecropsy cases in other hospitals which metthe same criteria were also included. Weanalysed the history of illness, CT scans ofthe brain, cerebral angiograms and necropsyfindings of these eight patients. Angiographicexamination of the cerebral arteries was per-formed on patients 1, 2, 5, and 7 after pro-dromal ischaemic neurological symptoms orbrain infarction. The patients, seven men andone woman, were aged 61-77 (mean 66)years. The risk factors of the patients wereassessed. All patients were hypertensive, andtwo had diabetes mellitus and another hadhypercholesterolaemia (table 1).

At necropsy, the brains were removed andfixed in 10% formalin. After fixation, the cir-cle of Willis and its major branches wereremoved and decalcified in 45% formic acid

17 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.17 on 1 January 1994. Dow

nloaded from

Ogata, Masuda, Yutani, Yamaguchi

Table I Clinical profiles of eight patients with large intracranial artery thrombosisRisk factors

Patient Agel Prodromal Survivalnumber Sex HT DM HC ischaemic symptoms Mode of onset Areas of infarct after infarct Causes of death

1 61/M + + - - Progressive R PICA area 19 days Respiratory failuredue to infarction

2 67/F + - - TIAs (2 times) and Abrupt LMCA 2 area 12 days PulmonaryRIND during embolismthe last 32 days

3 77/M + - - - Progressive RMCA 2 area 17 days Bleeding fromstomach cancer

4 55/M + - - - Progressive R PICA area 4 days Respiratory failuredue to infarction

5 66/M + + + Repeated TIAs Progressive L PICA area 12 days Respiratory failureduring the last 18 due to infarctiondays

6 73/M + - - - Progressive R ACA 2 area 17 days Intraperitonealbleeding

7 61/M + - - - Insidious RMCA 1 area 8 days Brain herniationduring sleep due to infarction

8 68/M + - - - Insidious RMCA 2 area 25 days Congestiveduring sleep heart failure

HT = hypertension, DM = diabetes mellitus, HC = hypercholesterolaemia, M = male, F = female, + = present, absent, R = right, L = left, TIA = transientischaemic attack, RIND = reversible ischaemic neurological deficit, MCA 1 = middle cerebral artery trunk, MCA 2 = cortical branch of middle cerebral artery,ACA 2 = cortical branch of anterior cerebral artery, PICA = posterior inferior cerebellar artery.

;A

Figure 1 Patient 1. A photomicrograph of a platelet-fibrin thrombus (arrow) formed on

plaque ulceration in right vertebral artery (VA). G, gruel of atheroma. Haematoxylin andeosin. Scale bar = 0O2 mm.

Figure 2 Patient 3. Plaque rupture (between asterisks)and an occlusive thrombus (T) in a cortical branch of rightmiddle cerebral artery (MCA). G, gruel. Elastin vanGieson stain. Scale bar = 05 mm.

in 10% sodium citrate solution for two days.Histological examinations were performed atapproximately 3 mm intervals with hema-toxylin and eosin staining and other conven-tional methods. Complete serial sections ofthe site of thrombotic occlusion or muralthrombus were made when necessary. Specialattention was paid to the sources of theemboli to the brain when examining theheart, ascending aorta, and extracranialcarotid and vertebral arteries.The clinical and pathological features of

patient 2 have been reported previously.'3

ResultsTable 1 summarises the clinical profiles of theeight patients with large intracranial arterythrombosis. Anticoagulation or antiplatelettherapy was not given to any patient before orafter development of brain infarction.Complete blood count, total protein, fibrino-gen, prothrombin time and partial thrombo-plastin time were normal in all patients. Nopatient had atrial fibrillation. No case showedspecific pathological changes in the heart andextracranial arteries which would be theembolic source to the brain. Patients 2 and 5showed prodromal ischaemic neurologicalsymptoms in the relevant arterial systems.

Histopathological study of the thrombosedarteries revealed the presence of occlusivethrombi in all eight cases, but the occlusivethrombi in two patients were considered to bethromboemboli dislodged from the muralthrombi observed in the proximal intracranialarteries. In patient 1, the occlusive thrombi inthe right posterior inferior cerebellar artery(PICA) were histologically similar to themural platelet-fibrin thrombus observed inthe right vertebral artery (VA) 1 cm proximalto the orifice of the occluded PICA. Thismural thrombus was localised shortly distal tothe severe stenosis and attached to the plaqueulceration where many foam cells infiltrated(fig 1). In patient 2, the occlusive throm-boemboli were observed in the corticalbranches of the left middle cerebral artery

18 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.17 on 1 January 1994. Dow

nloaded from

Cerebral artery thrombosis

..;\*:.t : ii> o< . p Q . . .e t XFI it||I>

Figure 3 Patient 5. Rupture (between asterisks) ofplaque (P) infiltrated with foam cells,andformation of mural platelet-fibrin thrombi (TI) and an occlusive erythrocyte-richthrombus (T2) in left VA. Haematoxylin and eosin. Scale bar = 0 1 mm.

Figure 4 Patient 6. Intraplaque haemorrhage and an occlusive thrombus (T) in a

cortical branch of right anterior cerebral artery. Masson's trichrome stain.Scale bar = 0-2 mm.

(MCA), and they were histologically similarto the mural platelet-fibrin thrombus attachedto the fibrosed wall of the left MCA trunkimmediately distal to the severe stenosis.The occlusive thrombi observed in the

remaining six cases were considered to haveformed in situ and were non-embolic.Rupture of atheromatous plaques was ob-served in patients 3, 4, and 5. A platelet-fibrin thrombus occluded the lumen at thesite of the rupture of a fibrous cap over thegruel in the right MCA in patient 3 (fig 2),and in the right VA in patient 4. The site ofthe rupture was heavily infiltrated with foamcells in this patient. The plaque rupture was

observed at the greatest stenosis and itsimmediately proximal portion. The left VA ofpatient 5 showed a platelet-fibrin thrombusattached to the site of plaque rupture and an

occlusive erythrocyte-rich thrombus. The siteof plaque rupture was heavily infiltrated withfoam cells (fig 3). The plaque ruptureoccurred unrelated to segmental stenosis.Cholesterol crystal emboli were found in theleptomeningeal branches of the bilateral pos-

terior cerebral arteries resulting in multiplesmall cortical infarcts. The cholesterol crystalemboli probably derived from the plaque rup-

ture in the left VA may have been responsiblefor the transient ischaemic attacks.

Rupture of the plaque was not observed atthe sites of thrombosis in patients 6, 7, and 8.In patient 6, intraplaque haemorrhage was

found in the fibrosed intima of a corticalbranch of the right anterior cerebral arterywhere a platelet-fibrin thrombus occluded thelumen at the greatest stenosis. The portion ofthe thrombus in touch with the plaque wallshowed advanced organisation comparedwith the rest of the thrombus (fig 4). Patient7 demonstrated an occlusive thrombus show-ing a layered structure of thrombus materialof differing age at the greatest stenosis and itsimmediately distal portion of the right MCAtrunk (fig 5). This thrombus may haveformed successively by repeated muraldeposits causing progressive luminal narrow-

ing. The oldest thrombus was attached to theplaque wall immediately distal to the greateststenosis, where numerous foam cell infiltratedthe innermost area of the plaque. In patient 8a platelet-fibrin thrombus occluded thelumen of a cortical branch of the right MCAat the greatest stenosis. The portion of thethrombus in touch with the plaque wallshowed advanced organisation comparedwith the rest of the thrombus (figure 6).

Figure 7 illustrates the stenosis of the

Table 2 Histopathologicalfindings of thrombosed large intracranial arteriesPatient Intraplaque Greatestnumber Thrombosed arteries Mural or occlusive Rupture ofplaque haemorrhage stenosis(%)1 R VA Mural - 982 LMCA 1 Mural - 993 R MCA 2 Occlusive + - 764 R VA Occlusive + - 775 L VA Occlusive + - 716 R ACA 2 Occlusive - + 977 RMCA 1 Occlusive - - 958 R MCA 2 Occlusive - - 93

R = right, L = left, VA = vertebral artery, MCA 1 = middle cerebral artery trunk, MCA 2 = cortical branch of middle cerebralartery, ACA 2 = cortical branch of anterior cerebral artery, + = present, -= absent. Greatest percent stenosis is expressed bypercentage of plaque occupying original luminal area.

19 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.1.17 on 1 January 1994. Dow

nloaded from

Ogata, Masuda, Yutani, Yamaguchi

Figure 5 Patient 7. Occlusion oflumen by platelet-fibrin thrombi (T2, T3) developingon an organising thrombus (Ti) in rightMCA trunk. Foam cells infiltrate innermost area

ofplaque. Haematoxylin and eosin. Scale bar = 0 1 mm.

..-....S^ .

Figure 6 Patient 8. An occlusive thrombus (T) in a cortical branch of right MCA.Masson's trichrome stain. Scale bar = 0-2 mm.

examined arteries and the localisation of thethrombi. All showed extensive stenosis of thelumen due to atherosclerotic changes and thestenosis rate at the point of maximal luminalreduction varied from 71 to 99% (88 ± 1 1%:mean ± SD) of the original luminal area

(table 2). The luminal diameters of the great-est stenosis were <09 mm in seven of eightpatients. Percentage stenosis of the luminaldiameter of the greatest stenosis comparedwith the luminal diameter 5-10 mm proxi-mal to it was 70 ± 28% (mean ± SD) (range:3-90%). Formation of thrombosis unrelatedto segmental stenosis of the arteries was

observed in only one patient. Thrombosis ofthe arteries therefore, occurred in relation tosegmental stenosis in most cases. The plaquerupture occurred at the greatest stenosis andits immediately proximal portion in two casesand at a site unrelated to segmental stenosisin one. Conversely, thrombus formationobserved in the remaining five patientsoccurred without plaque rupture, at thegreatest stenosis in two cases, and at sitesimmediately distal to the segmental stenosisin three.

DiscussionThe patients subjected to this study had riskfactors for atherosclerosis, such as hyperten-sion, diabetes mellitus, and hypercholestero-laemia. Other abnormalities predisposing'2 tothrombus formation, including blood coagu-lation disorders, collagen diseases, and vas-culitides, however, were not observed in anypatients. The histopathological findings ofthis study may therefore represent the cere-bral artery thrombosis that occurs in thosewith advanced atherosclerosis.The histopathological characteristics of

arterial segments with thrombus formationobserved in this study are: 1) rupture of ath-erosclerotic plaque, and 2) thrombus forma-tion without plaque rupture occurring at thegreatest stenosis or its immediately distal por-tion with or without intraplaque haemor-rhage.

In cases demonstrating rupture of theplaque, the essential histological features areexposure of connective tissue and other tissueelements and of clusters of lipid-ladenmacrophages and extracellular necrotic mate-rial of the gruel, to the arterial lumen. Thecontact of potent platelet activators derivedfrom the vascular tissue elements, and tissuefactor, a key protein in the activation of thecoagulation cascade, in the gruel and in foamcell-rich regions of the plaque'4 with blood inthe arterial lumen contributed to thrombusformation. This mechanism of thrombus for-mation in cerebral arteries was described byConstantinides.2 In our study, plaque rupturewas observed in three of eight cases.One patient in our study showed thrombus

formation over a fibrosed plaque walldemonstrating intraplaque haemorrhage.Intraplaque haemorrhage, not from the back-flow of blood through plaque rupture, hasbeen observed at the sites of thromboticocclusion of the cerebral arteries.'3 This ismainly due to the rupture of poorly supportedintimal small blood vessels, and is a possibletrigger mechanism for large intracranial arterythrombosis because it may release thrombo-genic substances into the lumen.

In our study, plaque rupture or intraplaquehaemorrhage was not essential for the throm-bus formation. In four cases, occlusive ormural thrombi developed at the greateststenosis or its immediately distal portionwithout plaque rupture or intraplaque haem-orrhage. Such thrombus formation has beenobserved in six of 39 thrombosed segments of

20

mm

on April 11, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.1.17 on 1 January 1994. D

ownloaded from

Cerebral artery thrombosis

Direction of blood flow

1J, 31 4 0.6 5 J1 30 .1 29

-~~~~~~~~~~----

Mural thrombus over plaque ulceration, Occlusive thrombus over plaque rupture,and its embolisation and embolisation of atheroma content

2 J1 1 6 41 02 6 1.18 -0-3

: ... .:::::::..S _~~~~~~...Mural thrombus on fibrosed intima, and Occlusive thrombus over fibrosed intimaits embolisation with intraplaque haemorrhage

3 1 220 0-02 7 - 2.7 4-08

.CTTT::

Occlusive thrombus over plaque rupture Occlusive thrombus with layered structureover fibrosed intima

to an occlusive one. Thus mural thrombi maygrow to form occluding thrombi as shown inpatient 7, and also may give rise to artery-to-artery embolism as shown in patients 1 and 2.Furthermore, advanced organisation of theportion of platelet-fibrin thrombus in touchwith plaque wall in patients 6 and 8 suggeststhat mural thrombi may grow to form occlu-sive thrombi.Thrombus leading to a myocardial infarc-

tion is generally formed at the site of severestenosis in coronary arteries,9 10 whereas thesite of thrombus formation in relation tostenosis in large intracranial arteries has notbeen previously reported. 1-3 The presentstudy has confirmed that thrombosis of thelarge intracranial arteries occurs at the site ofgreatest stenosis or immediately distal to it inmost cases, in the presence or absence ofplaque rupture or intraplaque haemorrhage.

These findings create a theoretical back-ground for antiplatelet, anticoagulation andthrombolytic therapy to prevent or treatatherothrombotic stroke, especially inpatients with segmental stenosis of the largeintracranial arteries demonstrated by cerebralangiography.

.... : Platelet-fibrin thrombus,: Erythrocyte-rich thrombust7:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,I~~~~~~~~~~~~~~~~~~~~~~~~

Figure 7 Schematic drawing of distmibution and localisation of atherosclerotic changesand thrombi in eight patients. The diameter at the greatest stenosis and 5-10 mm proximalto it are shown in millimeters.

cerebral arteries by Sadoshima et al,' in threeof 12 thrombosed segments of coronary arter-ies from patients with regional transmuralmyocardial infarction by Davies andThomas,8 and 1 1 of 51 coronary arterythrombi by Falk.10 A combination of dis-turbed blood flow at stenotic lesions, subse-quent endothelial damage and shear-inducedplatelet adhesion may contribute to thethrombus formation.'5 Furthermore, an accu-mulation of lipid-filled macrophages was pre-sent in such lesions, and may contribute tothe endothelial damage and exposure of tis-sue factor to the blood components.'1'6Our study has shown formation of non-

occlusive mural thrombi or occlusive thrombiwithout plaque rupture. The thrombilocalised at sites immediately distal to thegreatest stenosis in patient 7 demonstrated alamellated structure suggesting step-wise pro-gression and extension of the mural thrombi

We thank Drs Jun Karasawa and Ikuo Ihara for permission tostudy their necropsy cases.

1 Paterson JC. Capillary rupture with intimal hemorrhage inthe causation of cerebral vascular lesions. Arch Pathol1940;29:345-54.

2 Constantinides P. Pathogenesis of cerebral artery throm-bosis in man. Arch Neurol 1967;83:422-8.

3 Sadoshima S, Fukushima T, Tanaka K. Cerebral arterythrombosis and intramural hemorrhage. Stroke 1979;10:411-4.

4 Capron L. Extra- and intracranial atherosclerosis. InVinken PJ, Bruyn GW, Klawans HL, Toole JF, eds.Handbook of clinical neurology, vol 53. New York:Elsevier, 1988;91-106.

5 Constantinides P. Plaque fissures in human coronarythrombosis. J7 Atheroscler Res 1966;6:1-17.

6 Friedman M. The coronary thrombus: its origin and fate.Hum Pathol 1971;2:81-128.

7 Horie T, Sekiguchi M, Hirosawa K. Coronary thrombosisin pathogenesis of acute myocardial infarction. Histo-pathological study of coronary arteries in 108 necrop-sied cases using serial section. Br Heart Jf 1978;40:153-61.

8 Davies MJ, Thomas T. The pathological basis andmicroanatomy of occlusive coronary thrombus forma-tion in human coronary arteries. Philos Trans R Soc Lond[Biology] 1981;296:225-9.

9 Baroldi G. Diseases of the coronary arteries. In: SilverMD, ed. Cardiovascular pathology. New York: ChurchillLivingstone 1983:317-91.

10 Falk E. Plaque rupture with severe pre-existing stenosisprecipitating coronary thrombosis. Characteristics ofcoronary atherosclerotic plaques underlying fatal occlu-sive thrombi. BrHeartj 1983;50:127-34.

11 Davies MJ. Thomas AC. Plaque fissuring - the cause ofacute myocardial infarction, sudden ischaemic death,and crescendo angina. BrHeartj 1985;53:363-73.

12 Davies MJ. Thrombosis in acute myocardial infarctionand sudden death. Cardiovasc Clin 1987;18:151-9.

13 Masuda J, Ogata J, Yutani C, et al. Artery-to-arteryembolism from a thrombus formed in stenotic middlecerebral artery. Report of an autopsy case. Stroke 1987;18:680-4.

14 Wilcox JN, Smith KM, Schwartz SM, et al. Localizationof tissue factor in the normal vessel wall and in the ath-erosclerotic plaque. Proc Natl Acad Sci USA 1989;86:2839-43.

15 Mustard JF, Packham MA, Kinlough-Rathbone RL.Platelets, blood flow, and the vessel wall. Circulation1990;81(suppl I):I-24-7.

16 Masuda J, Ross R. Atherogenesis during low-level hyper-cholesterolemia in the nonhuman primate. I. Fattystreak forrnation. Arteriosclerosis 1990;10: 164-77.

4 - 2.2 4 0-7 8 4 1-9 4 0-9

Occlusive thrombus over plaque rupture Occlusive thrombus over fibrosed intima

//////2: Gruel of atheroma,| J;'] Intraplaque haemorrhage,

.440

21

on April 11, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.1.17 on 1 January 1994. D

ownloaded from