TIAs in the ED: A Pragmatic Approach Despite Imperfect Clinical Tools

Emergency Physicians are empowered to make one of the most expensive and risk-laden decisions in healthcare: admission to the hospital versus discharge. In the case of transient ischemic attacks (TIA), inappropriate discharge may result in a missed imminent stroke, whereas admitting all low-risk patients would contribute to unnecessary expense. Thus, appropriate risk stratification and clinical decision support tools are paramount to ensure optimal care of TIA patients in the emergency department (ED). While acknowledging the limitations of currently available tools, this AHA Scientific Statement offers a pragmatic approach to the diagnosis, workup, and risk reduction of TIA in the ED.

Making the diagnosis of TIA is a key step in identifying patients who are at risk for near term progression to acute stroke. The tools available to diagnose and treat stroke have evolved dramatically over the past decade or two. In the ED, we now routinely utilize imaging modalities such as diffusion weighted magnetic resonance imaging to identify infarcted tissue, computed tomography angiography (CTA) to identify clinically relevant large vessel atherosclerotic disease and acute large vessel occlusion, and CT perfusion imaging (CTP) to evaluate for potentially salvageable (via thrombolysis and/or thrombectomy) ischemic brain tissue. Despite these advancements, the recognition and initial management of TIA remains a challenge. While advanced imaging may help with risk stratification, TIA remains primarily a clinical diagnosis by frontline clinicians in the ED.

By definition, TIA patients must be back at their neurologic baseline. Asymptomatic upon ED arrival does not mean well, however. A parallel may be drawn to cases of syncope or chest pain where a patient is asymptomatic in the ED and has a reassuring examination but may be at high risk of a major adverse cardiac event if discharged. Obtaining the appropriate history and considering the diagnosis of TIA is at times straightforward. In other cases, patient factors and clinical presentations can make the diagnosis quite difficult. For example, patients' symptoms may not trigger clinicians to consider the possibility of cerebrovascular disease. This is especially the case in populations with low health literacy and in patients with whom there is a language barrier. Further, ascertaining an unfamiliar patient's "neurologic baseline" may be difficult. Ideally, a tool would be available to assist clinicians in identifying patients early in the triage process to facilitate the correct workup and management. Available stroke screening tools such as the Cincinnati Prehospital Stroke Scale or the NIH stroke scale (NIHSS) are not useful for TIAs since symptoms are resolved at the time of evaluation and the provider must rely primarily on a detailed history.

Eliciting anterior circulation symptoms such as unilateral extremity weakness, facial droop, and dysarthria by history allows easier recognition of these TIAs, while non-localizing symptoms of posterior circulation TIA such as transient dizziness or nausea make recognition of TIA in that setting very challenging. Indeed, up to half of patients who have a TIA resulting from vertebrobasilar (VB) ischemia may present with no focal neurological symptoms. Further, just as the NIHSS is weighted heavily to anterior circulation strokes, many of the validated scoring tools used to risk stratify TIAs favor anterior circulation findings such as unilateral weakness or sensory loss. Findings such as gait disturbance or visual symptoms are not included in the ABCD²/ABCD³/ABCD^3-I scores, which limits the utility of these scales in risk stratifying posterior circulation TIAs. Although the recently validated Canadian TIA score does include "History of gait disturbance" and "History of vertigo", the remainder of the tool is largely focused on anterior circulation symptoms. In that study of 7,607 ED TIA patients, only 1.4% of patients had a stroke within seven days. The area under the curve for the Canadian TIA Score was higher than that of the ABCD² (0.70 (95% confidence interval 0.66 to 0.73) versus 0.60 (0.55 to 0.64); location of stroke (anterior versus posterior) was not reported for the 108 patients who had a stroke within seven days. It was also unclear whether "history" of unilateral weakness, gait disturbance or vertigo in the Canadian TIA score referred to the presenting event or any history of these symptoms. The Canadian TIA score includes 13 variables, stratifies patients into low, medium and high risk and is not currently in widespread use.⁹ Other limitations of existing tools for TIA recognition and risk-stratification are well-discussed in this scientific statement. Nonetheless, the statement offers a useful approach to TIA in the ED as shown in the Figure. Availability of adequate imaging with expertise to interpret the images in real time is key for TIA diagnosis and workup. While the ABCD² remains the workhorse for TIA risk stratification and emergency physicians should be facile with its use, the Canadian TIA score may offer improved risk stratification if proven to be readily useful in the "real world" of ED TIA evaluation and management.

In conclusion, although there have been numerous advances in the diagnosis and treatment of acute stroke, the initial recognition, evaluation and management of TIA continues to rely heavily on the frontline clinician, frequently an emergency physician. A detailed clinical history and high index of suspicion for TIA are crucial to identifying at-risk patients. Availability of real-time advanced imaging in the ED with available expertise to interpret the images, and rapid outpatient follow up after ED discharge are essential for reducing risk while avoiding unnecessary admissions as appropriate. Notably, despite the advancements in stroke care available to many, underserved and rural communities still share the heaviest burden with limited access to both inpatient and outpatient stroke expertise. By providing remote access to an acute stroke expert through telemedicine technology, regional systems of stroke care should facilitate appropriate TIA evaluation and management for patients in these communities. Finally, future work is needed on easy to use, validated diagnostic and risk stratification tools to assist clinicians, especially for posterior circulation TIAs.