ORIGINAL 
Annals of Nuclear Medicine Vol. 7, No. 1, 21-27, 1993 
Factors causing prolonged hypoperfusion after transient ischemic attack 
Yoshinari ISAKA, Keiichi ASHIDA, Osamu IIJI and Masatoshi IMAIZUMI 
Department of Nuclear Medicine. Osaka National Hospital, Osaka, Japan 
Even during the symptom-free stages, patients with a TIA often experience cerebral blood flow disturbances. In order to evaluate the factors which cause this abnormality, we studied the cerebral blood flow disturbance, anatomy and clinical status in 21 patients after TIAs. The results of 99mTc-hexamethyl-propylene-amine oxime SPECT were compared with CT, cerebral angiogram, cerebrovascular risk factors and clinical findings to determine which factor is most responsible for the hypoperfusion of brain after TIA. The overall sensitivity rates in detecting a lesion were 67% in SPECT and 19% in CT. The hypoperfused area tended to be large in patients who had intracranial, severe stenotic, multiple, or hemodynamically significant arterial lesions on the ipsilateral side. No such relationships were found between other examinations. We conclude that hypoperfusion after TIA essentially refiects a continuous cerebral blood flow disturbance that can be attributed to atherosclerosis of the cerebral arteries, with subsequent embolic and/or hemodynamic cerebral ischemia, although there may be a variety of processes. 
Key words : 
TIA, cerebral blood flow, cerebral angiography, 99mTc-HM-PAO SPECT 
INTRODUCTION 
SEVERAL STUDIES of regional cerebral blood flow (rCBF) or CBF-single photon computed tomography (SPECT) have been performed in order to clarify the pathophysiology of transient ischemic attack (TIA),1-7 because TIA presents a brief focal neurologic deficit and occurs as a prelude to ischemic stroke in as few as 20% and as many as 80% of cases.8 However, previous studies of CBF after TIA have produced disparate results; normal rCBF,1,2 ,decreased rCBF3-5 or hyperperfusion with ischemia6 have been reported. More recently, using 99mTc-hexamethyl-propylene-amine oxime (HM-PAO) ,SPECT, Bogousslavsky et al.7 found a high incidence of early stroke in patients who had prolonged hypoperfusion after TIA. Thus, it remains to delineate in further studies the mechanisms underlying the prolonged hypoperfusion after TIA. 
We now report the results of 99mTc-HM-PAO SPECT in patients after TIAs. The results from CBF-SPECT were compared with computed tomography (CT), cerebral angiogram, cerebrovascular risk factors, and clinical parameters in an attempt to identify which factor is most responsible for the prolonged hypoperfusion after TIA. 
PATIENTS AND METHODS 
Patients 
Twenty-one patients admitted to Osaka National Hospital, Osaka, Japan between 1989 and 1990 had a TIA. TIAS Were defined according to the classification of the National Institute of Neurological Disorders and Stroke.9 The mean age of the patients was 57.2+-12.1 (+-SD) years (range, 32 to 75 years). Thirteen were male and 8 females. The project was reviewed and approved by the Committee on Studies Involving Human Beings of the Osaka National Hospital. Informed consent was obtained in each case. 
Clinical 
A complete clinical history was recorded and a neurologic examination was performed for each patient at the time of selection by the same neurologist. The neurologic deficits during the attack were classified in terms of physical signs: (I) motor dysfunction, (II) loss of vision, (III) sensory symptoms, (IV) aphasia, and (V) combination of symptoms I-IV. 
CBF- SPECT 
We used 99mTc-HM-PAO to assess the hypoperfusion of the brain. 99mTc-HM-PAO was prepared from a freeze-dried kit containing 0.5 mg of HM-PAO, 7.6 ug of stannous chloride, and 4.5 mg of sodium chloride in a vial (Amersham Medical Ltd., Tokyo, Japan). 99mTc-HM-PAO was reconstituted with 740 MBq (20 mCi) of 99mTc-sodium pertechnetate in 5 ml of solution at least 10 min before injecting the tracer via a 21-gauge cannula filled in with saline and free of blood into an antecubital vein. Image acquisition was started 5 min post injection. SPECT scanning was performed with a single-head rotating gamma camera fitted with a low-energy, high-resolution parallel-hole collimator and linked to a computer system (Toshiba GCA 901A). Sixty four views, 30 s frames collected over 360, were recorded into a 128 x 128 matrix format, with the subject's head immobilized. Transaxial slices were reconstructed from the prefiltered raw data (9-point weighted filter) with the aid of filtered backprojection algorisms using a ramp filter. Sorenson's techniquelo (u=0.12cm-1) was used for attenuation collection. No scatter correction was performed. Transaxial sections at 2.7 mm intervals were used for reconstructing 8.1 mm-thick computed images in planes parallel to the orbitomeatal line (OML). The resolution of this imaging system was measured as 19mm full width at half-maximum (FWHM) in the plane of the reconstructed transverse sections. 
We analyzed the SPECT images without correction of the HM-PAO back diffusion.11 The CBF images of the clinically responsible cerebral hemisphere were reviewed by independent observers and classified with respect to the spatial extent or the location of the hypoperfused areas. The slices which had the largest hypoperfused area were used for analysis. The severity of hypoperfusion was graded as follows: (I) normal, (II) mild abnormality, generally less than 2.0 cm in the largest diameter, (III) discrete area of hypoperfusion larger than 2.0 cm involving less than one lobe, and (IV) multilobe involvement (Fig. 1). The location was classified into three groups: (I) no hypoperfused area, (II) subcortical lesion, and (III) cortical lesion. Discrepancies in SPECT ratings between observers were resolved by consensus. The mean interval from the last attack to SPECT was 25.6+-13.3 days (range, 6-52 days). 
CT 
Conventional x-ray CT's of 1-cm layers between contiguous slices of the head without contrast enhancement were taken throughout the entire brain in planes parallel to the OML. The CTS Were graded according to the presence (group II) or absence (group I) of the hypodense area. In all patients, CT receded SPECT. 
Cerebral Angiography 
All patients underwent cerebral angiography. Intra-arterial digital subtraction angiography (IA-DSA)12 was performed in 18 patients less than 70 years old, and the remaining three, 70 years old or more, received intravenous (IV)-DSA. The cerebral angiograms were reviewed by a neuroradiologist who had no knowledge of the SPECT findings. The severity of angiographic abnormalities was categorized with respect to percentage stenosis, location, multiplicity and hemodynamic significance. All measurements were made directly on the radiographs. The percentage stenosis for each patient was calculated by measuring the luminal diameter at the narrowest point and dividing it by the normal luminal diameter of the vessel below the site of the stenosis. If no narrowing of the vessel was observed, it was called a 0% stenosis and occlusion of the vessel was a 100% stenosis. The percentage of stenosis was classified into following groups: (I) no stenosis, (II) <255, (III) >=25% and <50%, (IV) >=50% and <75%, and (V) >=75%.
The location of the vascular lesions was classified into the following four groups : (I) no lesion, (II) extracranial, (III) intracranial, and (IV) extracranial plus intracranial. 
The multiplicity of the lesions was classified into the three groups : (I) no vascular lesion, (II) solitary, and (III) multiple. 
Evidence of a hemodynamically significant stenosis13,14 was determined by the delayed filling of the middle cerebral artery branches relative to the external carotid branches (carotid), the prolonged crossfilling through the anterior circle of Willis (Willis), the arterial supply through the ipsilateral ophthalmic artery from the external carotid artery (ophthalmic), or the arterial supply through anastomotic channels across the surface from the anterior cerebral circulation (meningeal). The patients were classified into three groups: (I) no vascular lesion, (II) hemodynamically insignificant lesion, and (III) hemodynamically significant lesion. 
Cerebrovascular risk factors and other parameters The data included the following 12 information items : 
Risk factors. Age, sex, blood pressure measured by indirect auscultation on admission, current or former cigarette smoking, fasting blood glucose concentration and cholesterol concentration the day after admission, and venous hematocrit on admission. 
Cardiac investigations. Left ventricular hypertrophy (LVH), atrial fibrillation (AF), ischemic change in electrocardiography (ST) and two dimensional echocardiography (2DE). 
Other parameters. End tidal CO2 concentration (PETCO2) during the SPECT. 
Statistical analysis 
Comparisons of the ordinal ranks of each data item were done by the Kruskal-Wallis and Mann-Whitney U tests. 
RESULTS 
Clinical features 
Fourteen patients (66%) suffered from motor dysfunction ; the neurologic deficits in the other patients were a loss of vision in three (14%), aphasia in one (5%), sensory symptoms in one (5 %), and sensorimotor deficit in two (5 %).
The mean number of TIA episodes before the study was 3.2+-2.6; six patients (29 %) experienced only a single episode of TIA, 12 (57%) two to five episodes, and three (14%) more than five episodes. 
The mean duration of TIA was less than 5 min in 15 patients (71 %), less than 30 min in four (19 %), less than 5 hr in one (5 %), and less than 8 hr in one (5 %).
Hypoperfused area was found on the symptomatic hemisphere in 14 patients (67 %); seven patients (33%) were categorized as group I, eight (38%) as group II, and six (29 %) as group III. The hypoperfused area was most often located in the lenticular-capsular region (8 patients), followed in order of frequency by the parietal cortex (5), temporal cortex (3), and frontal cortex or occipital cortex (1) (Table 1). 
The sensitivity of CT in detecting a lesion was lower than that of SPECT (19 % versus 67 %)¥ The infarct was seen on CT in four patients (19%); internal capsule in two, and corona radiata or basal ganglia in one with the largest diameter less than 2.0 cm. Two patients had a hypodense area corresponding to the hypoperfusion on SPECT (Fig. 2). In cardiac investigation, LVH was found in three patients (14%), and ischemic change in one (5 %). No patient had AF or abnormality in 2DE. 
Fifteen patients (71 %) showed signs of a stenotic lesion on angiography; extracranial arterial lesions were found in eight patients (38%), intracranial lesions in five (24%) and intracranial plus extracranial lesions in two (9%). The degree of stenosis was of group I in six patients (29 %), group II in seven (33%), group III in three (14%), group IV in two (10%) and group V in three (14%). Four patients (19%) had a hemodynamically significant stenosis; collateral supply through the Willis was found in two patients, and through the carotid in two (Table 2). 
Comparison of SPECT with other examinations 
The SPECT class was not related to the classification of the neurological symptoms, delay between last attack and the examination, mean duration of attacks, the number of attacks, cerebrovascular risk factors, cardiac investigations, PETCO2, or the CT class Table 3. 
Overall, the SPECT class and classification of the angiography were highly related (Tables 2, 3). The severity rating of the SPECT tended to be high in patients who had severe stenotic lesions (x2= 12.8; p<0.005), intracranial arterial lesions (x2= 12.4; p<0.01), hemodynamically significant stenosis (x2=10.1 ; p<0.05), or multiple vascular lesions on the ipsilateral side (x2=9.0; p<0.05). The hypoperfused area in the cortical territories was more extensive than that in the subcortical territories (p< 0.01 ). Severely stenosed lesions (x2 = 12.8 ; p<0.005), intracranial lesions (x2=12.3; p<0.005), multiple lesions (x2=9.5 ; p<0.01) in angiography were more frequently found in patients who had cortical hypoperfusion (Fig. 3) than those with a subcortical hypoperfusion (Table 1). 
DISCUSSION 
We found that, even during the symptom-free stages, hypoperfusion of the brain can be seen in about two-thirds of patients with TIAs, and that the extent of the hypoperfused area was related to the presence of the intracranial, multiple, severely stenotic, or hemodynamically stenotic lesions in the ipsilateral large-vessels. 
TIA has been thought to be caused either by distal occlusion of cortical branches of the artery from embolic material originating in a carotid atherosclerosis, or by distal hemodynamic consequences resulting from reduced blood flow through the carotid with adequate collateral flow.15 Furthermore, cardiac arrhythmias16 and lacunar infarctions17 are becoming increasingly recognized as conditions which can lead to symptoms identical to those of large-vessel atherosclerotic disease. From the viewpoint of CBF, the hypoperfused area of ischemic stroke may include infarcted tissue, regions with diaschisis,18 with diffuse loss of small neurons with normal findings on CT19 or with reduced cerebral perfusion pressure.14 It is, therefore, important to clarify the pathogenesis of TIA in relation to the mechanisms for TIA onset, and the cerebral blood flow disturbance. 
The hypoperfused area was most often located in the lenticular-capsular region in our study population; two out of eight patients showed a small hypodense area on CT which corresponded to the hypoperfused area on SPECT and was indicative of the symptoms. These patients, therefore, may be categorized as having cerebral infarction with transient signs (CITS)20 and the hypoperfused area was considered to reflect infarcted tissue. Because of the low sensitivity of CT to detect small lacunar infarctions,21 more lesions could be detected by means of magnetic resonance imaging in the other six patients. In addition to lacunar infarction caused by small-vessel occlusion,17 emboli from an intra-arterial source22 or a cardiac source28 may cause cerebral ischemia in the subcortical territories. In previous studies. prevalence of carotid disease was found only in 5 % of asymptomatic patients over the age of 60 years,24 and 31 % of patients with lacunar infarction.25 We found carotid stenosis in five out of eight (63 %) patients with a hypoperfused area in the lenticularcapsular region ; the degree of stenosis was 0% in three patients, <25% in three, and 25-50% in two. No intracranial or multiple lesions were observed. Furthermore, no patient had a past history of systemic hypotension, severe carotid artery stenosis, or abnormalities in the cardiac investigations. Thus, the higher incidence of abnormal carotid angiography in these patients may indicate that small-vessel obstruction as well as emboli from the large-vessel are major causes of the CBF disturbance. 
In the group with a hypoperfused area in the cortex, all patients had an abnormal angiogram. In these patients, the severity rating of the vascular stenosis and hypoperfusion, and the incidence of the intracranial or multiple lesions were higher when compared to the lenticular-capsular group. As the size of the atherosclerotic plaque increases, it may reduce perfusion pressure or increase the risk of thromboembolism. Hypoperfused areas during the symptom-free stage of TIA may, therefore, reflect previous or continuing6 cerebral tissue damage induced by hemodynamic or embolic mechanisms, or both, resulting from the atheromatous vessels. We found marked hypoperfusion in patients with evidence of a hemodynamically significant stenosis. The hypoperfused area in patients with severe (>60%) atherosclerotic plaque26 may be further classified according to the cerebrovascular reserve capacity (CRC).27 Hypoperfusion with normal CRC indicates that decreased metabolic demand is linked to decreased CBF due to the ischemic tissue damage (matched low-perfusion), and hypoperfusion with decreased CRC27 indicates that compensatory vasodilation due to severe reduction in cerebral perfusion pressure (CPP) is insufficient to maintain CBF (misery perfusion; CBF reduction with increased oxygen extraction fraction). Powers et al.14 reported that the primary determinant of cerebral perfusion pressure and CBF is the adequacy of collateral circulation pathways. If the cerebrovascular reserve capacity is measured in our patients, the cause of hypoperfusion may be more clearly defined. We believe that hypoperfusion in carotid TIA reveals either ischemic tissue damage with matched low-perfusion or a miser-perfusion state. 
We postulate that the close relationship between the presence of intracranial vascular lesions and the extent of the hypoperfused area can be explained by a high incidence of severe atherosclerotic lesion in the intracranial portion of the internal carotid artery in Japanese,28 while in Americans, severe lesions are located more often in the extracranial portion of the internal or common carotid arteries. Wecksler et al.13 reported that most of the events in American patients with hemodynamically significant carotid siphon stenosis were TIAS or minor strokes. The prognosis of the carotid siphon stenosis should be evaluated in future with respect to racial differences. 
The mechanisms causing TIAS in patients with normal results remain unclear. In spite of extensive testing, we found no abnormality in three patients. A small carotid ulcer may be missed by angiography,29 and transient arrhythmias or cardiac abnormalities below the resolution of echocardiography could result in emboli.30 Because of the delay between the onset of TIA and the tests any of the hypoperfused areas may have been restored to normal by the time of the test. 
We conclude that hypoperfusion after TIA is a long-lasting, relatively fixed abnormality of the cerebral circulation, which is secondarily caused by atherosclerosis of the cerebral arteries, with subsequent embolic and/or hemodynamic cerebral ischemia. None of the hypoperfused areas can be considered as clinically insignificant because, in most patients after TIA, the hypoperfused area may reflect previous clinical and/or subclinical episodes20 of cerebral ischemia. 
ACKNOWLEDGMENTS 
We wish to thank Mr. Yoshihito Itoi and Mr. Mutsuo Shimoi for their technical assistance. 
REFERENCES 
1 . Skinhroj E, Hoedt-Rasrnussen K, Paulson OB, et al : Regional cerebral blood flow and its autoregulation in patients with transient focal cerebral ischemic attacks. Neurology 20 : 485-493 1970 
2. Hill TC, Holman BL, Lovett R, et al: Initial experience with SPECT (single-photon computerized tomography) of the brain using N-isopropyl I-123 p-iodoamphetamine. J Nucl Med 23 : 191-195, 1982 
3. Rees JE, Du Boulay GH, Bull JWD, et al : Regional cerebral blood-flow in transient ischemic attack. Lancet 2: 1210-1213, 1970 
4. Yonekura M, Austin G, Poll N, et al : Evaluation of cerebral blood flow in patients with transient ischemic attacks and minor stroke. Surg Neurol 15: 58-65, l 980 
5. Vorstrup S, Hemmingsen R, Henrikson L, et al : Regional cerebral blood flow in patients with transient ischemic attacks studied by xenon-133 inhalation and emission tomography. Stroke 14: 903-910, 1983 
6. Hartmann A : Prolonged disturbances of regional cerebral blood flow in transient ischemic attacks. Stroke 16: 932-939 1985 
7. Bogousslavsky J, Delaloye-Bischof A, Regli F, et al : Prolonged hypoperfusion and early stroke after transient ischemic attack. Stroke 21 : 40-46, 1990 
8. Barnett HJM : Medical treatment of transient ischemia ; In Cerebral Blood Flow, Volume 1. Wood JH (ed.): NY, McGraw-Hill, pp 593-603, 1987 
9. National Institute of Neurological Disorders and Stroke : Special report from the national institute of neurological disorders and stroke: Classification of cerebrovascular disease III. Stroke 21 : 637-675, 1990 
10. Sorenson JA : Qualitative measurement of radioactivity in vivo by whole body counting. In : Instrumentation in Nuclear Medicine, Volume 2. Hine JG, Sorenson JA (eds.) : NY. Academic Press, pp 311-324, 1977 
11 . Lassen NA, Anderson AR, Friberg L, et al : The retention of [99mTc]-d,l-HMPAO in the human brain after intracarotid bolus injection: A kinetic analysis. J Cereb Blood Flow Metab 8 : S13-S22, 1988 
12. Ackerman RH: Noninvasive diagnosis of carotid disease in the era of digital subtraction angiography. Neurol Clin 1: 263-278, 1983 
13. Wecksler LR, Kistler JP, Davis KR, et al: The prognosis of carotid siphon stenosis. Stroke 17: 714-718, 1986 
14. Powers WJ, Press GA, Grubb Jr. RL, et al : The effect of hemodynamically significant carotid artery disease on the hemodynamic status of the cerebral circulation. Ann Intern Med 106: 27-35 1987 
15. Millikan CH: The pathogenesis of transient focal cerebral ischemia. Circulation 32: 438-450 1965 
16. Bogousslavsky J, Hachinski VC, Boughner DR, et al : Clinical predictors of cardiac and arterial lesions in carotid transient ischemic attack. Arch Neurol 43: 224-233 1986 
17. Hankey G, Warlow CP : Lacunar transient ischaemic attacks : a clinical useful concept? Lancet 1: 335-338, 1991 
18. Feeney DM, Baron JC: Diaschisis. Stroke 5: 187-830 1986 
19. Powers WJ, Martin WRW, Herscovitch P, et al: Extracranial-intracranial bypass surgery : hemody-namic and metabolic effects. Neurology 34: 1168-1174 1984

20. Bogousslavsky J, Regli F : Cerebral infarction with transient signs (CITS) : Do TIA's correspond to small deep infarcts in internal carotid artery occlusion? Stroke 15: 536-539, 1984 
21. Awad I, Modic M, Little RJ, et al : Focal parenchymal lesions in transient ischemic attacks : Correlation of computed tomography and magnetic resonance imagin. Stroke 17: 399-404 1986 
22. Autret A, Pourcelot L, Saudeau D, et al: Stroke risk in patients with carotid stenosis. Lancet 1: 888-890, 1987 
23. Bogousslavsky J, Hachinski VC, Boughner DR, et al : Clinical predictors of cardiac and arterial lesions in carotid transient ischemic attacks. Arch Neurol 43 : 229-233, 1986 
24. Colgan MP, Strode GR, Sommer JD, et al : Prevalence of asymptomatic carotid disease : Results of duplex scanning in 348 unselected volunteers. J Vasc Surg 8: 674-678 1988 
25. Kappelle LJ, Koudstaal PJ, van Gijn J, et al : Carotid angiography in patients with lacunar infarction : A prospective study. Stroke 19: 1093-1096, 1988 
26. Deweese JA, May AG, Lipchick EO, et al : Anatomic and hemodynamic correlations in carotid artery stenosis. Stroke 1 : 149-157, 1970 
27. Yamashita T, Kashiwagi S, Nakano S, et al : The effect of EC-IC bypass surgery on resting cerebral blood flow and cerebrovascular reserve capacity studied with stable Xe-CT and acetazolamide test. Neuroradiology 33: 217-222, 1991 
28. Nishimaru K, McHenry LC Jr, Toole JF : Cerebral angiographic and clinical differences in carotid system transient ischemic attacks between American Caucasian and Japanese Patients. Stroke 15 : 56-59, 1984 
29. Bornmyr F, Jungquist G, Oliverrona H, et al : Ulceration of carotid bifurcation : A preliminary report on diagnostic problems. Acta Chir Scand 152: 499-501, 1986 
30. Di Pasquale GD, Andreoli A, Pinelli G, et al : Cerebral ischemia and asymptomatic coronary artery disease : A prospective study of 83 patients. Stroke 17: 1098-1101, 1986