Waimei Tai, MD

Bo. et. al. compares two prediction calculators for use in predicting near term stroke risk in patients who present with TIA. One is a purely clinical score (ABCD2) which incorporates age, history of diabetes, blood pressure, and clinical symptoms and duration of symptoms into the calculation. The second uses similar clinical metrics in addition 3 other signs (recurrent TIA within 7 days, ipsilateral vessel narrowing, DWI lesion). 

Patients were often moved from low to medium, or medium to high risk using the additional criteria in the ABCD3-I score. Those in the highest strata in the ABCD- I score had a 41% risk of stroke in 90 days while those in the lowest strata had no stroke at 90 days.

This study highlights and validates the higher positive predictive value of the modified ABCD2 score (ABCD3-I) and validates it in an ethnically different cohort than the previously retrospectively identified TIA cohorts.

Based on the new(-ish) TIA guidelines published by the AHA in 2009, using a tissue based definition, it makes sense that those with a DWI lesion have a higher risk of recurrent event. Similarly, having a prior episode of TIA also increases your risk of near term ischemic event (TIA begets TIA). Carotid stenosis has long been proven to contribute to a larger proportion of near term stroke risk. I think using imaging criteria to help identify patients at highest of near term events will aid in appropriate triaging of patients in a resource constrained environment.

I suggest that vessel imaging be used in the ED to triage patients into low, medium and high risk strata, and allow this to guide an outpatient evaluation and management of TIA (similar to other groups in Paris or Australia). To my knowledge, only a few centers in the US do this.  What is the practice in your area? What would you do if you had additional information in the ED? Do you routinely recommend vessel imaging to help triage patients?