J Neurol (2009) 256:639644 DOI 10.

1007/s00415-009-0150-9

ORIGINAL COMMUNICATION

Christian Weimar Jens Benemann Roman Huber Thomas Mieck Stephen Kaendler Steven Grieshammer Zaza Katsarava Hans-Christoph Diener on behalf of the German Stroke Study Collaboration*

Long-term mortality and risk of stroke after transient ischemic attack

A hospital-based cohort study

Received: 27 June 2008 Received in revised form: 6 October 2008 Accepted: 12 November 2008 Published online: 15 April 2009 C. Weimar, MD () J. Benemann, PhD Z. Katsarava, MD H.-C. Diener, MD Dept. of Neurology University of Duisburg-Essen Hufelandstr. 55 45122 Essen, Germany Tel.: +49-201/723-2495 Fax: +49-201/723-5176 E-Mail: stroke.med@uni-essen.de R. Huber, MD Dept. of Neurology University of Ulm Oberer Eselsberg 45 89081 Ulm, Germany T. Mieck, MD Dept. of Neurology Brgerhospital Stuttgart Tunzhofer Str. 1416 70191 Stuttgart, Germany S. Kaendler, MD Dept. of Neurology Klinikum Heidenheim Schlosshaustr. 100 89522 Heidenheim, Germany S. Grieshammer, MD Dept. of Neurology Stdt. Klinikum H. Braun Karl-Keil-Str. 35 08060 Zwickau, Germany * Participating Departments of Neurology (investigator): Ostalbklinikum Aalen (M. Heyden, MD), Klinikum Bernburg (M. Mller, MD), Krankenanstalten Gilead Bielefeld (C. Hagemeister, MD), Kranken-

haus Buchholz (K. Luckner, MD), University of Essen (C. Weimar, MD), University of Freiburg (C. Fritzsch, MD), University of Greifswald (A. Khaw, MD), University of Hannover (K. Weissenborn, MC), Klinikum Heidenheim (S. Kaendler, MD), University of Jena (C. Terborg, MD), Krankenhaus Kln-Mehrheim (U. Frost, MD), University of Leipzig (D. Michalski, MD), Landesklinik Lbben (C. Rhrig, MD), University of Magdeburg (M. Goertler, MD), Ruppiner Kliniken Neuruppin (G. Zindler, MD), University of Rostock (A. Kloth, MD), Brgerhospital Stuttgart (T. Mieck, MD), University of Ulm (R. Huber, MD), Heinrich-Braun-Krankenhaus Zwickau (S. Grieshammer MD).

Abstract Background Stroke and mortality rates in patients with transient ischemic attack (TIA) differ widely between communitybased studies and research cohorts. Our aim therefore was to provide a reliable estimate for TIA patients treated in German neurology departments with an acute stroke unit. Methods A total of 1951 consecutively admitted TIA patients were prospectively documented in 13 centers and 1480 (75.9 %) gave consent for long-term follow-up. During a mean follow-up of 23.4 months, we assessed recurrent cerebrovascular events and cause of death in 1448 patients via standardized telephone interview in-

cluding confirmation of endpoint events by the treating physician. Results Overall 94 patients (6.5 %) suffered a stroke and 118 patients (8.1 %) died, 21 due to stroke. The Kaplan-Meier estimate for stroke during the first year was 4.4 % (95 % CI 3.25.6 %) which corresponds to a relative risk of 9.5 (95 % CI 7.412.3) compared to the population-based stroke incidence in Germany. The annual rates after the first year were 2.2 % (95 % CI 1.72.7 %) for stroke and 3.2 % (95 % CI 2.73.8 %) for death. Independent predictors for stroke during follow-up were age and previous cerebrovascular events. The ABCD2 score did not provide any meaningful prediction of stroke risk at 90 days. Conclusion While the in-hospital risk of stroke was low, long-term stroke rates in our well-defined multicenter hospitalbased cohort were comparable to a large randomized trial. In patients with a well-established diagnosis of TIA, only age and previous cerebrovascular events seem to constitute independent predictors for stroke during long-term follow-up. Key words TIA transient ischemic attack prognosis recurrent stroke prediction
JON 3150

640

Introduction
Patients with a transient ischemic attack (TIA) have long been recognized as a high risk group for subsequent stroke and therefore constitute an important target for improving preventive strategies. In a recent review, 7-day risk of stroke following TIA varied between 0 % and 12.8 % [1]. Likewise, conflicting results can be found in the literature on long-term risk of stroke and mortality following TIA or minor stroke in population-based cohorts [2, 3] and consecutive hospital admissions [49]. However, most of these studies included only low patient numbers resulting in wide confidence intervals or date back to the 1970s or 1980s when secondary prevention was still limited and widely underused. As summarized in a review from community-based studies, the strongest predictor of stroke recurrence over 5 years is diabetes [10]. Other predictors of stroke recurrence include the well-known cardiovascular risk factors such as increasing age, previous TIA, atrial fibrillation, high alcohol consumption, hemorrhagic index stroke, and arterial hypertension [10]. The 7-point ABCD2 score has been validated for prediction of short-term risk of stroke after TIA and includes age (> 60 years = 1 point), blood pressure ( 140/90 mmHg = 1 point), clinical features of unilateral weakness (2 points) or speech disturbance without weakness (1 point), diabetes (1 point), and duration of symptoms ( 60 min = 2 points, 2059 min = 1 point) as independent predictors [11]. While this score has been tested in several outpatient settings, it remains uncertain if its use would result in better risk stratification in TIA patients admitted to German neurology departments with acute stroke units. We therefore aimed to assess the long-term risk of stroke and mortality in patients with TIA and to investigate independent predictors and risk stratification of subsequent stroke on the ABCD2 score.

menting center. Patients were asked to provide informed consent for participation in the long-term follow-up study. If consent was given, a central, structured follow-up interview via telephone (or paper questionnaire upon request) was performed biannually with screening for recurrent cerebrovascular events and assessment of functional disability scales (Barthel Index, modified Rankin Scale) or cause of death. In case of a suspected recurrent cerebrovascular event, confirmation from the treating general physician or hospital could be obtained for 92 % of stroke events and 90 % of recurrent TIA. Local citizen/death registries were checked if any patient could not be reached for follow-up and uncertain causes of death were confirmed on death certificates. The 19 centers listed in the appendix consecutively documented 3065 patients with a clinical diagnosis of TIA. To reduce selection bias, we only included 2022 patients from 13 centers where > 65 % of all patients had given informed consent for central follow-up. In addition, 67 patients were excluded with isolated amaurosis fugax and 4 with death due to other causes than stroke during the initial hospital stay. Of the remaining 1951 patients, 1480 (75.9 %) gave informed consent for long-term follow-up which could be obtained in 1448 patients (97.8 %), while 30 patients (2.0 %) later withdrew consent and 2 patients were lost to follow-up. Biannual follow-up was stopped due to lack of funding in June 2007 resulting in a mean follow-up duration of 23.4 (range 659) months. Patients without follow-up were significantly older (mean 69.3 vs. 67.6 years, p = 0.001), had a higher modified Rankin score at discharge (mean 0.55 vs. 0.41, p = 0.009) and more often had TIA symptoms lasting 1 hour (70.7 % vs. 64.5 %, p = 0.012). No significant differences were found with regard to baseline severity of neurological deficits, recurrent cerebral ischemic events during hospital stay, cardiovascular comorbidity or risk factors. Statistics Baseline characteristics of all TIA patients were stratified by recurrent stroke and reported as percentages or mean (median). Statistical comparisons were carried out via Fishers exact test, Persons chisquare or Mann-Whitney U test for categorical and continuous variables, as appropriate. Independent predictors of time to stroke were identified by Cox-regression analysis including all baseline variables with significant differences upon univariate analysis. Smoking was not included in the final model because of an implausibly lower rate of recurrent events in smokers (due to a strong interaction with age). The Cox regression model is presented with hazard ratios, 95 % confidence intervals and p values. To evaluate the performance of the ABCD2 score, we calculated the area under the curve (AUC) by Receiver Operating Curve statistic. An AUC of 0.5 indicates no discrimination, and an AUC of 1.0 indicates perfect discrimination. All statistical analyses were performed in SPSS 14.0.

Methods
Most acute TIA patients who present to tertiary care centers in Germany are admitted to an acute stroke unit. However, we did not exclude patients who were admitted to a general ward in the participating centers. A diagnosis of TIA was given by the treating neurologist if acute neurological deficits of presumably vascular (ischemic) cause completely resolved within 24 h regardless of infarction shown on cerebral imaging [12]. Between 09/2002 and 12/2006, consecutive patients with TIA were recruited prospectively by 19 participating neurology departments with acute stroke unit (not all centers participated during the whole study-period, mean time of recruitment 31 months) on standardized case report forms which included demographic information, risk factors, comorbidities, neurological deficits, presumed etiology on the modified Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification [13], premorbid modified Rankin Scale (mRS) and mRS at discharge, results of diagnostic work-up, length of hospital stay and secondary prevention at discharge. After central double key stroke input of all baseline data, data quality was monitored via automated consistency checks and queries were sent to the docu-

Results
The 1951 documented TIA patients were 17 to 96 years old (mean 68.0 13.6 SD) and 929 (47.6 %) were women. 1756 patients (90.8 %) were treated on an acute stroke unit (median length of stay 2 days). Delay between symptom onset and hospital admission was < 3 h in 44.6 %, 3 6 h in 26.1 %, 624 h in 21.2 % and > 24 h in 8.1 % of the patients. A previous stroke was reported by 362 patients (18.6 %) and a previous TIA by 255 patients (13.1 %). Mean duration of neurological symptoms was < 1 hour in 642 patients (32.9 %) and 1 to 24 hours in 1251 patients (64.1 %). No information on symptom duration was available in 58 patients (3.0 %). Cerebral im-

641

aging was documented in 99.5 % (MRI in 61.2 %) and an acute infarction could be shown in 9.1 %. Mean length of hospital stay was 6.55 days (SD 4.72). Stroke etiology was classified as large artery disease in 12.1 %, small vessel disease in 31.7 %, cardioembolic in 19.7 % and other or undetermined etiology in 36.5 % of patients. At discharge, 892 patients (45.7 %) were on aspirin, 223 (11.3 %) were on phenprocoumon or warfarin, 355 (18.2 %) on clopidogrel, 217 (11.1 %) on aspirin/dipyridamol, 103 (5.3 %) on aspirin/clopidogrel, 2 (0.1 %) on ticlopidin, 91 (4.7 %) on high dose heparin and 5 (0.3 %) on various study medications (no antithrombotic medication or only low dose, low molecular weight heparin was documented in 63 (3.2 %) patients. Of 1448 patients with follow-up information, 58 patients (4.0 %) underwent revascularization of a high-grade carotid artery stenosis and 118 (8.1 %) died during follow-up, 21 (1.5 %) of cerebral stroke, 43 (3.0 %) of other cardiovascular causes, 42 (2.9 %) of non-cardiovascular causes and 12 (0.8 %) of unknown causes. The Kaplan-Meier estimate for death during the first year was 4.4 % (95 % CI 3.25.6 %) and the annual mortality after the first year was 3.2 % (95 % CI 2.73.8 %). During the entire follow-up period, a cerebral stroke (excluding recurrent TIA) occurred in 94 patients (6.5 %). Nine stroke events (0.6 %) occurred during the initial hospital stay in the documenting center. Respective stroke rates at 30 and 90 days were 14 (1.0 %) and 22 (1.5 %), respectively. The Kaplan-Meier estimate for recurrent stroke during the first year after index TIA was 4.4 % (95 % CI 3.25.6 %) and the annual stroke rate after the first year was 2.2 % (95 % CI 1.7 2.7 %). The relative risk of stroke compared to a German population-based stroke incidence study [14] in all TIA patients is displayed in Table 1. Upon univariate analysis, increasing age, non-smoking, previous cerebrovascular events, coronary artery disease, atrial fibrillation, and increasing modified Rankin Scale at discharge were associated with a significantly increased rate of stroke (Table 2). Cox regression analysis identified only age and previous cerebrovascular events as independent predictors for time to stroke (Table 3). Upon stratification on

the ABCD2 score, the 90-day stroke rate was 0.6 % in the low-risk group (03 points, 24.3 %), 1.8 % in the medium risk group (45 points, 46.8 %), and 1.5 % in the high risk group (67 points, 28.9 %), which was not significant. In addition, 41 patients (2.8 %) had a recurrent TIA during the initial hospital stay in the participating center and 82 patients (5.7 %) reported TIA during follow-up. Thus, 217 patients reported at least one cerebrovascular event including TIA. The Kaplan-Meier estimate for recurrent cerebrovascular events (including TIA) during the first year was 11.0 % (95 % CI 9.212.8 %) and the annual rate after the first year was 3.9 % (95 % CI 3.44.5 %). No significant difference in the unadjusted rate of stroke or recurrent cerebrovascular events including TIA could be found in patients with symptom duration < 1 hour vs. 124 hours (even after exclusion of patients with previous cerebrovascular events) although a significantly higher rate of recurrent TIA was observed in patients with symptom duration < 1 hour vs. 124 hours (11.8 % vs. 7.7 %, p = 0.012).

Discussion
Our data indicate a low stroke rate (0.6 %) during the acute hospital stay and a cumulative stroke risk of 4.4 % during the first year following hospital admission for acute TIA. These findings are relevant for in-hospital monitoring and patient information and moreover could guide the design of future clinical trials in secondary prevention. Like in our own previous study [15], we found a high proportion (64.1 %) of TIA patients with symptoms lasting 124 h. Patients with a short-lasting TIA may not recognize their symptoms or may be more likely to seek medical attention with a greater delay which may not result in hospital admission although patients with acute TIA in Germany are generally admitted to the hospital even if neurological deficits are completely restituted. Because patients in our study were indexed to the time of their initial admission to the hospital rather than

Table 1 Stroke risk after TIA stratified by age compared to a population-based stroke incidence study [14] Age 1844 4554 5564 6574 7584 85+ total N 105 127 288 478 358 92 1448 Events 2 5 10 32 35 10 94 Annual stroke rate first year* (%) 1.9 2.5 2.2 4.3 6.1 11.8 4.4 Expected stroke rate (%) 0.006 0.063 0.147 0.408 0.888 1.153 0.463 RR (95 % CI) first year 305.7 39.8 14.9 10.5 6.9 10.2 9.5 (7.4-12.3) Annual stroke rate after first year* (%) 0 0.8 1.1 2.2 5.4 1.5 2.2 RR (95 % CI) after first year 12.7 6.8 5.4 6.1 1.3 4.8 (3.3-6.8)

* Kaplan-Meier-estimate for stroke risk

642

Table 2 Baseline characteristics of all TIA patients with complete follow-up and stratified by recurrent stroke

Recurrent stroke all (n = 1448)* Mean age (median), years Women, % Premorbid modified Rankin Scale (> 2), % Previous cerebral ischemia, % Previous stroke, % Previous TIA only, % Diabetes mellitus, % Arterial hypertension, % Hypercholesterolemia, % Coronary artery disease or MI, % Peripheral arterial disease, % atrial fibrillation, % Smoking during past five years, % alcohol abuse TOAST classification large artery atherosclerosis cardiac embolism small vessel disease other/undetermined Cerebral MRI imaging, % Acute infarction on imaging, % Aphasia, % Hemiparesis, % Symptom duration < 10 min, % 1 h, % modified Rankin Scale at discharge, mean Antithrombotic medication at discharge Antiplatelets Anticoagulation None/unknown 67.6 (69) 46.3 5.1 32.5 19.1 13.4 22.8 70.8 33.8 20.3 5.5 13.8 16.9 4.6 12.9 20.2 31.8 35.0 62.0 9.4 27.4 54.3 6.2 64.5 0.41 81.4 15.0 3.7 no (n = 1354) 67.2 46.5 4.9 31.0 17.9 13.0 22.4 70.3 34.0 19.7 5.2 13.2 17.5 4.6 12.6 20.2 31.8 35.4 63.0 9.3 26.7 54.2 6.1 64.1 0.39 81.5 14.8 3.6 yes (n = 94) 72.7 43.6 8.0 53.8 35.1 18.3 28.7 78.7 30.6 28.7 9.8 21.3 9.0 5.6 16.9 21.7 32.5 28.9 46.8 10.8 38.3 55.3 6.5 70.7 0.73 78.7 17.0 4.3 p-value < 0.001 0.595 0.208 < 0.001 < 0.001 0.156 0.163 0.100 0.556 0.046 0.091 0.042 0.040 0.602 0.546

0.002 0.585 0.017 0.915 0.823 0.217 < 0.001 0.794

* > 2 % data missing for previous modified Rankin Scale (4.0 %), TOAST classification (5.2 %) hypercholesterolemia (2.8 %) and symptom duration (2.6 %)

Table 3 Cox regression analysis for stroke in all TIA patients with follow-up and complete information on all variables (N = 1423) Cox regression Hazard ratio Age (per year) Previous cerebrovascular event Atrial fibrillation Coronary artery disease or MI Modified Rankin scale at discharge Aphasia 1.033 1.982 1.337 1.190 1.139 1.405 95 %-CI 1.013 1.305 0.800 0.749 0.953 0.916 1.054 3.011 2.236 1.892 1.361 2.156 p-value 0.001 0.001 0.268 0.462 0.153 0.119

to the time of onset of TIA, patients may not have been admitted for their TIA but for a subsequent stroke shortly after TIA. Of 8396 patients admitted with acute stroke in the same centers and recruitment period, about 1 % reported a preceding TIA within 48 h of stroke onset (data not shown). Our in-hospital short-term risk therefore is hardly comparable to other outpatient studies which found a substantially higher short-term rate of stroke following TIA [1]. Nevertheless, a recent population-based sequential comparison could show a striking decrease in the 90-day rate of stroke from 10.3 % in a primary care setting versus 2.1 % for urgent assessment and immediate treatment in a specialized outpatient clinic [16]. Like in the French SOS-TIA study [17], the 90-day stroke risk predicted on the ABCD2 score in our cohort was about 5-fold higher than the observed rate.

643

Thus, in addition to case-mix and methodological differences [18], the low stroke rate in our hospital-based study might be at least partly due to early diagnostic work-up and initiation of appropriate secondary prevention strategies [19]. We could not confirm the predictive value of an acute infarction on cerebral imaging, as recently proposed in smaller TIA cohorts [2022]. Likewise, duration of TIA was not a significant predictor for stroke in our large, well-defined, hospital-based cohort. Thus, the predictive value of imaging and symptom duration may reflect the accuracy of diagnosis rather than a different prognosis in TIA patients. Due to the low event rate, the ABCD2 score did not show any significant risk prediction at 90 days. When comparing the long-term rates of stroke and mortality, one has to keep in mind that our hospitalbased cohort was selected especially with regard to younger age compared to other population-based TIA studies [2, 23, 24]. Age in turn was confirmed as an independent predictor for stroke during follow-up. Our longterm results are comparable with a 10-year follow-up study of patients with TIA or minor ischemic stroke from the Dutch TIA Trial (recruitment 198689; mean age 65 years 10.1 SD; arterial cause of cerebral ischemia)

which found a cumulative risk of first recurrent stroke of 4.7 % at 1 year and 12.0 % at 5 years after the index event [25]. Compared to age-stratified stroke incidence rates in the general German population [14], the annual risk of stroke still was increased about 10-fold. Shortcomings of our study include a follow-up rate of only about 75 % of all eligible patients and a telephone follow-up which was based on a structured interview screening for cerebrovascular events. This may have resulted in an underestimation of recurrent cerebrovascular events. On the other hand, the diagnosis of the index TIA was well established through hospital stay in a stroke center and medical confirmation was sought for all reports of recurrent cerebrovascular events. In conclusion, our incidence rates for stroke and mortality can be regarded as representative for TIA patients admitted to neurology departments with acute stroke units and could guide health care decisions and design of future clinical trials of secondary prevention.
Conflict of interest The authors declare no conflict of interest. Acknowledgments This study was supported by the German Research Foundation (Deutsche Forschungsgemeinschaft, DI 327/8-1, DI 327/8-2). The funding sources had no involvement in the study. We thank P. Dommes, PhD, for central data collection and management.

References
1. Giles MF, Rothwell PM (2007) Risk of stroke early after transient ischaemic attack: a systematic review and metaanalysis. Lancet Neurol 6:10631072 2. Terent A (1989) Survival after stroke and transient ischemic attacks during the 1970s and 1980s. Stroke 20: 13201326 3. Dennis M, Bamford J, Sandercock P, Warlow C (1990) Prognosis of transient ischemic attacks in the Oxfordshire Community Stroke Project. Stroke 21:848853 4. Heyman A, Wilkinson WE, Hurwitz BJ, Haynes CS, Utley CM, Rosati RA, Burch JG, Gore TB (1984) Risk of ischemic heart disease in patients with TIA. Neurology 34:626630 5. Hankey GJ, Slattery JM, Warlow CP (1991) The prognosis of hospitalreferred transient ischaemic attacks. J Neurol Neurosurg Psychiatr 54: 793802 6. Howard G, Evans GW, Crouse JR 3rd, Toole JF, Ryu JE, Tegeler C, FryePierson J, Mitchell E, Sanders L (1994) A prospective reevaluation of transient ischemic attacks as a risk factor for death and fatal or nonfatal cardiovascular events. Stroke 25:342345 7. Clark TG, Murphy MF, Rothwell PM (2003) Long term risks of stroke, myocardial infarction, and vascular death in low risk patients with a non-recent transient ischaemic attack. J Neurol Neurosurg Psychiatry 74: 577580 8. Carolei A, Candelise L, Fiorelli M, Francucci BM, Motolese M, Fieschi C (1992) Long-term prognosis of transient ischemic attacks and reversible ischemic neurologic deficits: a hospital-based study. Cerebrovasc Dis 2:266272 9. Muuronen A, Kaste M (1982) Outcome of 314 patients with transient ischemic attacks. Stroke 13:2431 10. Hankey GJ (2003) Long-term outcome after ischaemic stroke/transient ischaemic attack. Cerebrovasc Dis 16(Suppl 1):1419 11. Johnston SC, Rothwell PM, NguyenHuynh MN, Giles MF, Elkins JS, Bernstein AL, Sidney S (2007) Validation and refinement of scores to predict very early stroke risk after transient ischaemic attack. Lancet 369:283292 12. Special report from the National Institute of Neurological Disorders and Stroke. Classification of cerebrovascular diseases III. (1990) Stroke 21: 637676 13. Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE (1993) Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 24(1):3541 14. Kolominsky-Rabas PL, Weber M, Gefeller O, Neundoerfer B, Heuschmann PU (2001) Epidemiology of ischemic stroke subtypes according to TOAST criteria: incidence, recurrence, and long-term survival in ischemic stroke subtypes: a population-based study. Stroke 32:27352740 15. Weimar C, Kraywinkel K, Rodl J, Hippe A, Harms L, Kloth A, Diener HC, German Stroke Data Bank Collaborators (2002) Etiology, duration, and prognosis of transient ischemic attacks: an analysis from the German Stroke Data Bank. Arch Neurol 59:15841588 16. Rothwell PM, Giles MF, Chandratheva A, Marquardt L, Geraghty O, Redgrave JN, Lovelock CE, Binney LE, Bull LM, Cuthbertson FC, Welch SJ, Bosch S, Carasco-Alexander F, Silver LE, Gutnikov SA, Mehta Z (2007) Effect of urgent treatment of transient ischaemic attack and minor stroke on early recurrent stroke (EXPRESS study): a prospective population-based sequential comparison. Lancet 370:14321442

644

17. Lavallee PC, Meseguer E, Abboud H, Cabrejo L, Olivot JM, Simon O, Mazighi M, Nifle C, Niclot P, Lapergue B, Klein IF, Brochet E, Steg PG, Leseche G, Labreuche J, Touboul PJ, Amarenco P (2007) A transient ischaemic attack clinic with round-the-clock access (SOS-TIA): feasibility and effects. Lancet Neurol 6:953960 18. Hankey GJ, Dennis MS, Slattery JM, Warlow CP (1993) Why is the outcome of transient ischaemic attacks different in different groups of patients? BMJ 306:11071111 19. Hackam DG, Spence JD (2007) Combining multiple approaches for the secondary prevention of vascular events after stroke: a quantitative modeling study. Stroke 38:18811885

20. Prabhakaran S, Chong JY, Sacco RL (2007) Impact of abnormal diffusionweighted imaging results on shortterm outcome following transient ischemic attack. Arch Neurol 64: 11051109 21. Sciolla R, Melis F (2008) Rapid identification of high-risk transient ischemic attacks: prospective validation of the ABCD score. Stroke 39:297302 22. Coutts SB, Eliasziw M, Hill MD, Scott JN, Subramaniam S, Buchan AM, Demchuk AM (2008) An improved scoring system for identifying patients at high risk of stroke and functional impairment after an acute transient ischemic attack or minor stroke. Int J Stroke 3:310

23. Dennis M, Bamford J, Sandercock P, Warlow C (1991) The prognosis of transient ischemic attacks in the community. The Oxfordshire Community Stroke project. Stroke 21:848853 24. Ricci S, Cantisani AT, Righetti E, Duca E, Spizzichino L (1998) Long-term follow-up of TIAs: the SEPIVAC study. Neuroepidemiology 17:33 25. van Wijk I, Kappelle LJ, van Gijn J, Koudstaal PJ, Franke CL, Vermeulen M, Gorter JW, Algra A (2005) Long-term survival and vascular event risk after transient ischaemic attack or minor ischaemic stroke: a cohort study. Lancet 365:20982104

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.