Optimizing the risk estimation after a transient ischaemic attack - the ABCDE⊕ score

Background and purpose: The risk of stroke after a transient ischaemic attack (TIA) can be predicted by scores incorporating age, blood pressure, clinical features, duration (ABCD-score), and diabetes (ABCD2-score). However, some patients have strokes despite a low predicted risk according to these scores. We designed the ABCDE+ score by adding the variables 'etiology' and ischaemic lesion visible on diffusion-weighted imaging (DWI) -'DWI-positivity'- to the ABCD-score. We hypothesized that this refinement increases the predictability of recurrent ischaemic events.

Methods: We performed a prospective cohort study amongst all consecutive TIA patients in a university hospital emergency department. Area under the computed receiver-operating curves (AUCs) were used to compare the predictive values of the scores with regard to the outcome stroke or recurrent TIA within 90 days.

Results: Amongst 248 patients, 33 (13.3%, 95%-CI 9.3-18.2%) had a stroke (n = 13) or a recurrent TIA (n = 20). Patients with recurrent ischaemic events more often had large-artery atherosclerosis as the cause for TIA (46% vs. 14%, P < 0.001) and positive DWI (61% vs. 35%; P = 0.01) compared with patients without recurrent events. Patients with and those without events did not differ with regard to age, clinical symptoms, duration, blood pressure, risk factors, and stroke preventive treatment. The comparison of AUCs [95%CI] showed superiority of the ABCDE+ score (0.67[0.55-0.75]) compared to the ABCD(2) -score (0.48[0.37-0.58]; P = 0.04) and a trend toward superiority compared to the ABCD-score (0.50[0.40-0.61]; P = 0.07).

Conclusion: In TIA patients, the addition of the variables 'etiology' and 'DWI-positivity' to the ABCD-score seems to enhance the predictability of subsequent cerebral ischaemic events.