The duration of symptoms in transient ischemic attack

Methods.

We reviewed the clinical records of 81 consecutive patients with carotid TIAs who were admitted to our hospital between June 1990 and December 1996 within 2 weeks of TIA onset. Patients consisted of 64 men and 17 women age 65.8 ± 9.9 years. Carotid TIA was diagnosed if the patient had a brief episode of unilateral motor dysfunction (i.e., weakness, paralysis, or clumsiness of extremities or face) or isolated aphasia that resolved within 24 hours based on the definition of the National Institute of Neurological Disorders and Stroke Classification III.5 Patients with amaurosis fugax were excluded from the current study.

Information about the approximate duration of a TIA was obtained from patients or their families in all cases. The symptom duration was reported as follows: <15 minutes in 23 patients, ≥15 and <30 minutes in 11, ≥30 and <60 minutes in 7, ≥1 and <2 hours in 9, ≥2 and <12 hours in 17, and ≥12 hours in the remaining 14 patients. The patients were then classified into three groups for further analysis: 23 patients with TIAs <15 minutes (Group 1); 27 patients with TIAs ≥15 minutes and <2 hours (Group 2); and 31 patients with TIAs ≥2 hours (Group 3).

The following clinical characteristics were analyzed with respect to the duration of TIAs: 1) patient age and gender; 2) history of brain infarction, TIA, myocardial infarction, and definite angina pectoris; 3) risk factors for stroke, including hypertension, diabetes mellitus, hypercholesterolemia, cigarette smoking, and alcohol consumption; 4) use of antiplatelet agents or anticoagulants at onset of TIA; 5) potential cardiac sources of emboli; 6) significant arterial pathologies in the ipsilateral carotid system; 7) number of TIAs; and 8) evidence of recent brain infarction on CT or MRI.

The following risk factors were identified: 1) use of antihypertensive agents, systolic blood pressure >160 mm Hg or diastolic blood pressure >95 mm Hg on admission for hypertension; 2) use of oral hypoglycemic agents, insulin, or glycosylated hemoglobin (HbA1C) >6.4% for diabetes mellitus; and 3) use of antihyperlipidemic agents or serum cholesterol level >220 mg/dL for hypercholesterolemia.

To detect potential cardiac sources of emboli (emboligenic cardiac diseases), all patients were examined by 12-lead EKG, 24-hour EKG monitoring, and transthoracic echocardiography. Emboligenic cardiac diseases included nonvalvular atrial fibrillation (NVAF), acute myocardial infarction, old myocardial infarction with intraventricular thrombus, mitral valve disease, prosthetic cardiac valve, implantation of pacemaker, dilated cardiomyopathy, and atrial septal defect (ASD), which may cause paradoxical embolism.

We carried out color-flow duplex carotid ultrasonography, conventional angiography, and MR angiography (MRA) to evaluate significant arterial pathologies in the ipsilateral carotid system. Color-flow duplex carotid ultrasonography (Toshiba SSA 270A or 260A, Toshiba Inc., Tokyo, Japan) was performed in all patients. The carotid arteries were examined in longitudinal and transverse planes from the anterior, lateral, and posterior approaches using B-mode imaging and color-flow imaging, allowing direct measurement of the residual lumen diameter of a stenotic lesion. Conventional cerebral angiography was performed in 10 patients (12%), MRA in 39 patients (48%), and both were performed in 23 patients (28%). Thus, a total of 72 patients (89%) was examined for the presence of arterial diseases in the ipsilateral carotid system. The grade of stenosis of the internal carotid artery (ICA) was determined by the method used in the North American Symptomatic Carotid Endarterectomy Trial.8 The percent of stenosis of the middle cerebral artery (MCA) was similarly calculated by measuring the diameter of the stenotic lesion and of the adjacent intact portion. The lesions were considered significant if arterial stenosis was >50% or if an ulceration was evident in the carotid bifurcation.

CT was performed in all patients within 4 weeks of onset of the TIA to exclude nonischemic brain lesions such as brain hemorrhage, chronic subdural hematoma, and brain tumors, and to detect relevant infarcts. A recent infarct was defined as a hypodense lesion in the vascular territory corresponding to the patient’s symptoms.

T1-weighted (repetition time [TR]/echo time [TE], 500/20) and T2-weighted (TR/TE, 1,800/110) MRI using a 1.5-T superconducting magnet was carried out within 40 days of TIA onset in 58 patients (71%). T1-weighted MRI with Gd-DTPA injection (0.1 mmol/kg, Schering, Germany) was performed in 41 patients. Ischemic lesions were considered to be recent if they were enhanced with Gd-DTPA or if they were hyperintense on T2-weighted images and isointense to slightly hypointense on T1-weighted images, which were located in the vascular region corresponding to the patient’s symptoms. The lesions on CT or MRI were classified as cortical infarcts, lacunar infarcts (small, deep lesions less than 15 mm in diameter), and border zone infarcts (lesions located between two adjacent arterial territories).

The risk factors, arterial or cardiac diseases, and CT and MRI findings were compared among the three groups using the χ² test. The number of TIAs was analyzed in patients with emboligenic cardiac and arterial disease using the Mann-Whitney U test. When a significant difference was observed, we applied sensitivity-specificity curve analysis to obtain a cutoff value for the duration of TIAs. Based on this assessment, TIA patients were defined as having either long-duration or short-duration TIAs. Statistical analysis was performed using a commercially available software package (Stat-View, version 4.5; SAS Institute, Cary, NC). Data are expressed as mean ± SD. The p values <0.05 were considered statistically significant.

Results.

A history of cerebral infarction was found more frequently in Groups 2 and 3 than it was in Group 1. No other differences in history, risk factors, or number of TIAs were observed among the three groups (table 1). Emboligenic cardiac diseases were detected in 24 patients (29%): NVAF in 13 patients, prosthetic mitral valve in 6, pacemaker in 2, acute myocardial infarction in 1, old myocardial infarction with intraventricular thrombus in 1, and ASD in 1. Anticoagulant was administered to eight patients: five with a prosthetic mitral valve, one with NVAF, and two with ischemic heart disease. An antiplatelet agent had been administered to 20 patients to treat previous TIA (n = 7), previous cerebral infarction (n = 4), ischemic heart diseases (n = 6), and other conditions (n = 2).

Color-flow duplex carotid ultrasonography revealed significant arterial diseases ipsilateral to the affected side in 18 patients (22%). Significant arterial diseases were found in the carotid axis ipsilateral to the affected side in 25 (51%) of 49 patients examined by conventional angiography. The arterial lesions consisted of the following: MCA occlusion in 2 patients, stenosis ≥50% in the horizontal portion of the MCA in 8, ICA occlusion in 2, ICA stenosis ≥50% in 10, and ICA stenosis <50% with ulceration in 3. MRA was performed in 23 patients who were not examined by conventional angiography. Significant arterial lesions were observed in three patients: a stenosis of 75% at the M1 segment in one patient, ICA stenosis of 90% in the ICA in one, and ICA occlusion in one.

Table 2 summarizes the emboligenic cardiac or arterial diseases found in the 72 patients who underwent cardiac and arterial examinations. Patients in Groups 2 and 3 had emboligenic cardiac or arterial diseases more frequently than patients in Group 1 (p < 0.05).

We performed a sensitivity-specificity curve analysis between concomitant emboligenic cardiac or arterial diseases and TIA duration. The results indicated that the cutoff time point for dividing the TIA patients into two groups was 1 hour (figure). Twenty-nine of 35 patients (83%) with long-duration TIA had either accompanying emboligenic cardiac disease or significant arterial disease significantly more frequently than short-duration TIA patients (17 of 37 patients [46%], p < 0.001) (table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Sensitivity (thick line) and specificity (thin line) of TIA duration to predict the presence of emboligenic cardiac and arterial diseases. The cutoff point for discriminating the long-duration TIA from the short-duration TIA was 1 hour.

CT revealed recent infarcts in only seven patients (8%). Fifty-five of the 74 patients (74%) in whom CT revealed no recent infarcts underwent MRI analysis, which demonstrated recent infarcts in 14 patients. Combining CT and MRI results, 21 patients had recent infarcts. All lesions but one were located in the cerebral cortex and the remaining one in the deep structure. Of the 21 patients, 15 (71%) had the duration longer than 1 hour, and 18 (86%) had emboligenic cardiac or arterial diseases. Of the 62 patients who underwent MRI studies or had recent lesions on CT, recent infarcts were present in 15 of 33 patients (45%) with TIA lasting ≥1 hour but in only 6 of the 29 patients (21%) with a TIA duration <1 hour (p < 0.05) (see table 3).

The number of TIAs was more frequent in the 30 patients with arterial diseases (median 1, 25th percentile 1, 75th percentile 4, range 1 to 9) than it was in the 18 patients with emboligenic cardiac diseases (median 1, 25th percentile 1, 75th percentile 2.7, range 1 to 3) (p < 0.05).

Discussion.

Waxman and Toole3 reported that some patients with TIAs had a CT-proven relevant cerebral infarct. They referred to this condition as a cerebral infarction with transient signs. TIA patients with a relevant infarct on CT tend to have a longer duration of attack than those without an infarct.3,5 An association of carotid stenosis or carotid ulcerated plaque and cerebral infarction on CT has been suggested.9,10 After the development of MRI, Fazekas et al.11 also reported that the presence of recent infarcts is associated with a higher incidence of significant cardiovascular disorders. We observed a similar relationship between the duration of TIAs and the presence of recent infarcts on CT or MRI.

The current study demonstrates that the duration of TIA has a significant relationship with the presence of arterial or cardiac abnormalities. Emboligenic cardiac or arterial diseases were seen more frequently in patients with long-duration TIAs (≥1 hour) than they were in patients with short-duration TIAs. Short- and long-duration TIAs appear to have different mechanisms. We believe that long-duration TIAs are usually caused by transient occlusion of the large vessels with emboli from the heart or proximal arteries. Thus, these cases are often accompanied by cerebral infarcts on CT or MRI. Short-duration TIAs, however, are caused by microembolic embolisms or small perforating vessel disease. At present, TIAs are diagnosed based on clinical history without considering CT or MRI lesions. If TIAs are defined as a transient focal ischemic episode without infarction, the criterion for time duration of TIA is at most 1 hour rather than 24 hours.

There are several problems associated with the current study. The indicated duration of TIAs was not always accurate because this information depends on the memories of patients or their families and on MRI or enhanced MRI, and angiography was not performed in all patients. A prospective study is needed to confirm the current results. Recently, the importance of other emboligenic factors for cerebral embolism, such as aortic arch atheromatous plaques, patent foramen ovale, and atrial septal aneurysms, has been emphasized, and these conditions are more sensitively detected by transesophageal echocardiography (TEE) than by transthoracic echocardiography.12 However, the current patients were not routinely evaluated by TEE, and thus the above risk factors may have been overlooked.

Fisher13 emphasized the importance of TIA as a risk factor for cerebral infarction. Based on the high incidence of stroke after a TIA (4 to 8% during the first month, 12 to 13% during the first year, and 24 to 29% during the next 5 years14), prevention of cerebral infarction following TIA is important. Toole14 recently stated that TIAs represent a marker rather than a risk factor for cerebral infarction that frequently follows TIA. As such, particular attention should be paid to the cerebrovascular risk factors and to emboligenic cardiac and arterial diseases to prevent cerebral infarction following a TIA. According to our results, patients with long-duration TIA should be more closely examined for the presence of cardiac and arterial diseases than those with short-duration TIA.