From clinical to tissue-based dual TIA: Validation and refinement of ABCD3-I score.

Qiliang Dai,Wen Sun,Yunyun Xiong,Graeme J Hankey,Lulu Xiao,Wusheng Zhu,Minmin Ma,Wenhua Liu,Dezhi Liu,Qiankun Cai,Yunfei Han,Lihui Duan,Xiangliang Chen,Gelin Xu,Xinfeng Liu
DOI: https://doi.org/10.1212/WNL.0000000000001444
IF: 9.9
2015-01-01
Neurology
Abstract:OBJECTIVE:To investigate whether dual tissue-defined ischemic attacks, defined as multiple diffusion-weighted imaging lesions of different age and/or arterial territory (dual DWI), are an independent and stronger predictor of 90-day stroke than dual clinical TIAs (dual TIA). METHODS:Consecutive patients with clinically defined TIA were enrolled and assessed clinically and by MRI within 3 days. The predictive ability of the ABCD clinical factors, dual TIA, and dual DWI was evaluated by means of multivariate logistic regression. RESULTS:Among 658 patients who were included in the study and completed 90 days of follow-up, a total of 70 patients (10.6%) experienced subsequent stroke by 90 days. Multivariate logistic regression indicated that dual DWI was an independent predictor for subsequent stroke (odds ratio 4.64, 95% confidence interval 2.15-10.01), while dual TIA was not (odds ratio 1.18, 95% confidence interval 0.69-2.01). C statistics was higher when the item of dual TIA in ABCD3-I score was replaced by dual DWI (0.759 vs 0.729, p = 0.035). The net reclassification value for 90-day stroke risk was also improved (continuous net reclassification improvement 0.301, p = 0.017). CONCLUSION:Dual DWI independently predicted future stroke in patients with TIA. A new ABCD3-I score with dual DWI instead of dual clinical TIA may improve risk stratification for early stroke risk after TIA.
What problem does this paper attempt to address?