Risk prediction of subsequent early stroke in patients with transient ischemic attacks.

BACKGROUND: Prognostic clinical scores (ABCD2 and ABCD3-I), as well as specific clinical signs and symptoms (e.g. fluctuations) have been used to predict early stroke risk in patients admitted to hospital after transient ischemic attacks (TIAs). We compared the utility of these scores and signs for prognosis and outcome.

METHODS: 235 patients with TIAs admitted to our Comprehensive Stroke Center entered the study. Patients were monitored over 3 days with detailed brain imaging [diffusion-weighted imaging (DWI) was performed either directly on admission or within 24 h from admission and was repeated in cases of stroke recurrence], vascular ultrasound imaging, repeated neurological scoring and continuous ECG monitoring. Duration, fluctuation of symptoms, clinical patterns of initial signs and/or symptoms as well as general patient characteristics and stroke risk factors, including atrial fibrillation (AF), were analyzed and recorded in our stroke databank. Fluctuation of symptoms was defined as the complete remitting and relapsing of TIA symptoms for ≥2 times in the acute phase within the first 24 h. This differs from the 'dual TIA' definition of the ABCD3-I score, which is defined as 'an earlier TIA within 7 days of the index event'. ABCD2 and ABCD3-I scores were calculated and the patients were placed into three categories: 'low', 'moderate' and 'high' risk for every score. Risk associations were assessed by the χ(2) test and the φ-coefficient.

RESULTS: Out of 235 patients, 17 patients (7.2%) experienced an early stroke during hospitalization (mean duration 7.4 ± 2.7 days). ABCD2 scores failed to predict early stroke (p = 0.544). ABCD3-I scores correlated better with early stroke recurrences (p = 0.021). Positive DWI findings (6/17; 35.3%), presence of carotid stenosis (3/17; 17.6%) or AF (1/17; 5.9%) alone failed to predict early stroke. Fluctuations of symptoms, however, showed a significant prediction for early stroke after TIA: 13/17 (76.5%) patients (p < 0.001). The combination of symptom fluctuation and MR-DWI-positive findings (4/17; 23.5%) also turned out to be statistically significant in this regard (p = 0.003), while the combination of symptom fluctuations with carotid stenosis ≥50% did not (p = 0.151). Combining fluctuations with carotid stenosis and DWI-positive findings did not improve the result (p = 0.029).

CONCLUSIONS: While the ABCD3-I score is indeed very useful, symptom fluctuations seem to be the best available and an easily accessible and applicable parameter for individual prediction of a high early stroke risk after TIAs.