Article Text


Transient ischaemic attack preceding anterior circulation infarction is independently associated with favourable outcome
  1. M Sitzer1,
  2. C Foerch1,
  3. T Neumann-Haefelin1,
  4. H Steinmetz1,
  5. B Misselwitz2,
  6. C Kugler3,
  7. T Back4
  1. 1Department of Neurology, Johann Wolfgang Goethe-University, Frankfurt, Germany
  2. 2Geschaeftsstelle Qualitaetssicherung Hessen, Eschborn, Germany
  3. 3Institut fuer Integrative Versorgung in der Medizin (IVM), Gießen, Germany
  4. 4Department of Neurology, Philipps University Marburg, Marburg, Germany
  1. Correspondence to:
 Dr M Sitzer;

Statistics from

Experimental studies in various animal models have provided evidence that short episodes of global or focal ischaemia partially protect the brain against subsequent ischaemic damage. The term “ischaemic tolerance” has been established for this phenomenon, both in the cerebral circulation and in other organs (see recent review by Kirino1). There are only a few clinical reports supporting the notion that ischaemic tolerance may also exist in stroke patients who have had a short episode of focal cerebral ischaemia—that is, a transient ischaemic attack (TIA) prior to cerebral infarction.2,3 The purpose of the present study was to investigate the association between the outcome of anterior circulation infarction and prior TIA in stroke patients from a large scale multicentre stroke registry.

The ASH (Arbeitsgruppe Schlaganfall Hessen) database is a prospective hospital based stroke registry for the federal state of Hesse in Germany. This registry covers all stroke units (nine neurological hospitals) in this state with the obligation to include all patients presenting with cerebral ischaemia or intracerebral haemorrhage. Thus, ~90% of stroke patients treated in these hospitals from 1998 to 2000 are included in the ASH database with 12 571 complete datasets (2960 with TIA, 8484 with cerebral infarction, 1127 with intracerebral haemorrhage). Patients with proven anterior circulation infarction (computed tomography (CT) or magnetic resonance imaging (MRI) scan) were included in the present analysis (n = 4969); from these, those patients presenting with stupor or coma were excluded (n = 172, 4.8% with previous TIA). The characteristics of the study population are presented in table 1. In 332 (6.9%) patients a history of TIA (according to previously published guidelines4) preceded final infarction. Favourable outcome was defined as a Modified Rankin Scale (MRS) score of ⩽1 and Barthel Index ⩾90 at discharge.

Table 1

Characteristics of stroke patients with proven anterior circulation infarction from the ASH stroke database

On univariate logistic regression analysis (inclusion method), we found that the odds ratio (OR) for a history of TIA preceding a stroke with a favourable outcome was 1.57 (95% confidence interval (CI) 1.26 to 1.97; p<0.001). The OR remained unchanged after the adjustment for age, sex, cause of the stroke (see table 1), risk factors, time between symptom onset and hospital admission, duration of hospital stay, medical centre, and secondary prevention treatment (1.58, 1.25 to 2.02; p<0.001). The inclusion of the MRS score at hospital admission reduced the adjusted OR to 1.52 (1.10 to 2.13; p = 0.012), although it was still significant. This association was also not altered after the exclusion of all patients with severe speech disturbances (1.77, 1.20 to 2.68; p<0.005). In a subset of patients (n = 1601), the European Stroke Scale (ESS) was determined at admission; the median ESS score was significantly higher in patients with preceding TIA (84 v 78; p<0.05, Mann–Whitney U test). On arbitrarily defining an ESS score ⩾70 as a mild neurological deficit, we found an adjusted OR of 1.56 (1.05 to 2.41; p<0.029) for a history of TIA preceding a mild neurological deficit at hospital admission. Furthermore, in the group of stroke patients (n = 1520) presenting with anterior circulation TIA at hospital admission, the prevalence of preceding TIA was 15.6% (n = 237). On including the TIA patients into the regression analysis, a preceding TIA had an OR of 2.76 (2.31 to 3.38; p<0.001) for a subsequent transient neurological deficit—that is, TIA and not a stroke, after the adjustment for the above mentioned confounders.

From the present analysis we conclude that TIA preceding cerebral anterior circulation infarction is independently associated with a less severe neurological deficit at hospital admission (as quantified by the ESS), and with a higher probability of a favourable outcome after hospital treatment. Furthermore, after a TIA the probability of a subsequent TIA instead of persisting stroke is 2.75 fold higher. Our findings support previous reports. Moncayo et al reported an adjusted OR of 1.98 (1.27 to 3.08) for a favourable outcome in stroke patients with preceding TIA after one month from a prospective single centre database.3 In a retrospective analysis, Weih and coworkers found an OR of 3.57 (1.39 to 9.14) for an independent state in stroke patients with preceding TIA at discharge.2 Thus, our data fit well within this context and support a robust association between a preceding TIA and a better outcome after subsequent cerebral ischaemia.

The main limitation of our analysis is that detailed information about the TIAs, such as the duration, the vascular territory in which the TIA occurred, and the time interval between the TIA and the final infarction was not available. Nevertheless, on the basis of the size of the database, its prospective design, and the inclusion of data from several centres, the results are reliable and make it unlikely that the positive association found (and previously described) is only due to hitherto unidentified bias.

Whether this association reflects “ischaemic tolerance” in humans could not be answered in the present study or from the literature.1–3 There are at least two other mechanisms which may influence the severity of stroke in this subset of patients. Consistent with previous reports, we found a higher percentage of large vessel atherosclerosis as the presumed cause of stroke in the group with preceding TIA (see table 1).2,3 First, in these patients, embolus size or composition may differ significantly from emboli from other sources and may lead to more distal vessel occlusion, earlier recanalisation, and smaller infarcts.5 Secondly, the capacity of collateral intracranial pathways may differ in patients and may preserve brain tissue from infarction.6 Observational studies are unable to control for these factors and thus cannot quantify the real impact of ischaemic tolerance in stroke patients. However, one observation in the present analysis suggests a neuroprotective effect: preceding TIA was significantly associated with a 1.52 fold higher probability of favourable short term outcome even after the correction for the degree of disability at hospital admission—that is, MRS score, suggesting improved recovery from initial neurological deficit.

In conclusion, our results support an independent association between preceding TIA and favourable outcome in subsequent cerebral ischaemia. Whether this association reflects inducible tolerance of brain tissue against ischaemic damage remains unanswered and should be elucidated in further studies taking the above mentioned considerations into account.


The following neurological hospitals contributed to the ASH database used for the present analysis: Asklepios Neurologische Klinik Bad Salzhausen (Prof Dr von Reutern), Horst-Schmidt-Kliniken Wiesbaden (Prof Dr Weisner), Johann Wolfgang Goethe-Universität Frankfurt am Main (Prof Dr Steinmetz), Klinikum Darmstadt (Prof Dr Claus), Klinikum Fulda (Prof Dr Langohr), Klinikum Kassel (Prof Dr Ferbert), Klinikum Weilmünster (Prof Dr Hornig), Krankenhaus Nordwest Frankfurt am Main (Prof Dr Janzen), Philips Universität Marburg (Prof Dr Oertel).


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.