Challenges in the diagnosis and management of transient ischemic attack

It is well established that transient ischemic attack (TIA) is a strong predictor of future ischemic stroke [1–3]. The 90-day stroke risk after TIA is estimated to be 10.5% with, notably, half of the risk occurring within 2 days following TIA [3]. While these data stress the need for urgency in TIA evaluation and accurate risk stratification, real-world clinical decision-making can be very challenging and complex. Indeed, making an accurate diagnosis of TIA, calculating stroke risk and initiating appropriate management are all important steps over which the treating clinician may stumble. Since diagnosis of TIA represents a golden opportunity to prevent subsequent stroke, it is critical that such errors are minimized.

Art of clinical diagnosis

Classically, TIA has been defined as a ‘sudden, focal neurologic deficit that lasts for less than 24 h, is presumed to be of vascular origin and is confined to an area of the brain or eye perfused by a specific artery’ [4]. Since symptoms have already resolved at the time of presentation in the majority of cases, objective examination findings are often lacking. The clinician must rely on history from a patient who may, in some instances, be unable to recall exactly what transpired. Vague non-medical descriptions of symptoms (e.g., heaviness’ or ‘confusion’) further compound the problem. Accurate diagnosis, therefore, depends heavily on the patient’s knowledge of symptoms and signs, their ability to communicate them effectively and the physician’s knowledge and ability to extract this vital information by asking the appropriate questions.

Surveys suggest that knowledge of TIA symptoms is quite poor among patients and physicians [5,6]. Among general practitioners, only 22% correctly identified all five typical TIA symptoms (i.e., focal weakness, focal sensory loss, aphasia, dysarthria and visual loss) and knew that TIA was defined as lasting less than 24 h (44% thought TIA could last up to 48 h and 9% thought even > 1 week) [5]. Despite clear differences in diagnostic accuracy among neurologists and non-neurologists [7], only 37% routinely consulted a neurologist for evaluation of suspected TIA patients [5]. Overall, studies assessing the accuracy of TIA diagnosis suggest that 30–62% of patients diagnosed with TIA have an alternative diagnosis on expert evaluation by a neurologist [7–12].

What clinical signs and symptoms can physicians use to identify TIA patients correctly? In a recent study of emergency-room diagnoses of TIA, the proportion of patients presenting with focal neurologic deficits who were misdiagnosed as having TIA was 60% when reviewed by stroke neurologists [12]. Three clinical features were useful in distinguishing TIA from common mimics: rapid onset within 1 min, lack of prior stereotyped transient neurologic attacks and absence of associated nonspecific or nonfocal symptoms. In addition, specific symptoms, such as isolated visual or sensory presentations, lack of carotid bruit and absence of focal signs on initial examination, may be clues to non-ischemic mimics, such as migraine or seizure [9–11,13]. Another clinical tool that incorporates clinical features into a validated score is a method using age, blood pressure, clinical symptoms, duration and diabetes (ABCD²; see later). While this was initially developed and validated as a prognostic tool, it has utility in distinguishing true TIA from mimics, such that higher scores were more likely in true TIA patients, as confirmed by expert neurologists [14].

Besides risk factors, rate of symptom onset and the type of symptoms, the duration of symptoms is also a critical piece of information in the acute evaluation of TIA. According to the classical definition, a TIA may last up to 24 h. However, it is well established that the majority of TIAs last less than 1 h and 90% resolve by 4 h [15]. Stated another way, only 14% of patients with symptoms lasting 1 h or greater completely recovered at 24 h [16]. Thus, the 24-h time limit is arbitrary and inaccurate in most cases. This may have significant treatment implications in the era of effective but time-sensitive thrombolytic therapies, such as tissue plasminogen activator (tPA), intra-arterial thrombolysis and mechanical thrombectomy. Armed with this arbitrary time-based definition, the treating physician is left to wonder whether a patient’s neurologic symptoms will resolve spontaneously within 24 h (TIA) or not (ischemic stroke). In the former scenario, interventions would be unnecessary, whereas in the latter, they would be indicated. This uncertainty might lead to withholding acute revascularization therapies for eligible patients in the false hope that symptoms will resolve within 24 h.

The emergence of modern neuroimaging has also improved diagnostic accuracy for TIA. Using acutely performed diffusion-weighted imaging (DWI), up to a third of patients with TIA symptoms lasting less than 1 h have evidence of cytotoxic damage on DWI, increasing to over 70% in those with symptoms lasting 12–24 h [17]. This diagnostic tool provides incontrovertible objective evidence of brain ischemia and improves diagnostic certainty that a TIA occurred. As such, DWI is known as the ‘troponin of stroke’.

Overall, these data unmistakably demonstrate the challenges in diagnosing TIA in real-world settings. While clinical tools such as the ABCD² score may be useful, their ‘cook-book’ utilization, especially by non-neurologists, may overlook some TIA patients at a high risk for stroke and misdiagnose others who do not have a TIA at all. For example, would not an episode of hemiparesis lasting less than 10 min in a 50-year-old man with atrial fibrillation undoubtedly be a reason for concern despite an ABCD² score of 0? Diagnostic certainty can only be enhanced by combining knowledge about clinical features (e.g., risk factors, type of symptoms, nature of onset, duration and ABCD² score) and neuroimaging markers (DWI) with sound clinical judgement and experience.

Calculating stroke risk following TIA

In true cases of reversible brain ischemia, TIA may be likened to ‘unstable angina’, heralding a vulnerable atherothrombotic state that can recur as a disabling ischemic stroke. Much in the way that cardiologists stratify coronary risk using troponin and thrombolysis in myocardial infarction score, clinical and diagnostic tools can be utilized to identify high-risk TIA patients. There are four main approaches to stroke-risk stratification after TIA: clinical, imaging-based, etiology-based and combined approaches. Of the clinical scores, the most validated and applied system is the ABCD² score, which uses five criteria to create a 0–7 point system: age (1 point for 60 years or older), blood pressure on first assessment (1 point for systolic greater than 140 or diastolic greater than 90 mmHg), clinical features (2 points for unilateral weakness and 1 point for speech impairment without weakness), duration of TIA (2 points for greater than 60 min and 1 point if between 10–59 min) and diabetes (1 point if present). When dichotomized, scores less than 4 were associated with a 2-day stroke risk of 1% and 90-day stroke risk of 3% compared with 5 and 12%, respectively, for scores of at least 4 [18].

The imaging-based approach utilizes DWI and magnetic resonance angiography (MRA) to identify high-risk TIA patients. The risk of stroke following DWI-positive TIA is estimated at 8.3–14.8% within the first week, increasing to as high as 32.6% at 90 days. By comparison, there seems to be a much lower risk of stroke following DWI-negative TIA (only 2.0% during hospitalization and 4.3% at 90 days) [19–21]. Furthermore, the combination of DWI lesions and intracranial large-artery occlusion found in TIA patients presenting within 24 h of symptom onset was associated with a 90-day stroke risk of 40.6% compared with 4.0% in those without DWI and MRA abnormality [21]. The observation that DWI-negative TIA is associated with a low risk of stroke might be due to the inclusion of nonischemic conditions, such as migraine or seizure misdiagnosed as true TIA, affirming the utility of DWI as a diagnostic test.

Combining clinical and radiographic features, one study found that those with symptoms lasting for longer than than 1 h, positive DWI, and evidence of large artery disease were at the highest risk of subsequent cerebral ischemic and vascular events [21]. While there is strong correlation between high-risk clinical features of TIA and DWI abnormality [22–24], the addition of DWI has been shown to increase prognostic value independently of the clinical scores [25].

The etiology-based approach presupposes that high-risk conditions, such as large-artery stenosis or atrial fibrillation, are predictors of subsequent stroke risk. Studies have confirmed the utility of this approach [20,26]. Biomarkers and modern neuroimaging may shed further light on risk stratification. For example, a significant proportion of acute TIA patients have perfusion deficits without DWI lesions or have large perfusion-diffusion mismatch, suggesting tissue at risk of subsequent infarction [27–29]. These latter approaches seem to address pathophysiology more squarely and may lend themselves to targeted therapies, such as acute revascularization of occlusive lesions, blood pressure augmentation and/or anti-thrombotic agents for specific conditions, such as atrial fibrillation.

Akin to unstable angina and myocardial infarction risk [30], future scoring tools for TIA should incorporate multiple clinical and radiographic parameters that can improve diagnostic accuracy, elucidate pathophysiology and prognosticate stroke risk in the effort to triage patients with TIA more effectively. Currently, a combined risk stratification strategy using first clinical then, if appropriate, neuroimaging criteria, followed by etiologic determination may be the optimal and most scientific approach. Given the complexity involved, consultation with a stroke neurologist in the acute evaluation of TIA patients seems warranted.

Acute management of TIA

Once a patient is diagnosed with TIA and then identified as ‘high risk’, what steps can be taken to lower stroke risk? Ischemic stroke patients treated with intravenous tPA within 3 h of symptom onset are 30% more likely to have minimal or no disability at 3 months [31]. Given the narrow time window for tPA administration and the high risk of stroke within 2 days after TIA, should we hospitalize all patients with TIA diagnosis in an observation unit to monitor for recurrent stroke symptoms?

One study looked at the cost–effectiveness of hospital admission for TIA on the sole basis of the potential benefits of administering tPA if symptoms recurred. Admission for all patients with TIA was borderline cost effective, while admission of high-risk patients (defined as patients with 2-day risk greater than 5%) was very cost effective [32]. Other benefits of hospitalization, such as starting preventative treatments, inpatient cardiac monitoring, and opportunity for stroke educations, were not considered in this study, suggesting an even greater rationale for hospitalization after TIA.

The current standard treatments in the management of TIA include antithrombotic therapy, blood pressure-lowering agents and statins. These therapies have been shown to prevent long-term risk of stroke, but the short-term effects in TIA have only recently been studied. The Effect of Urgent Treatment of Transient Ischemic Attack and Minor Stroke on Early Recurrent Stroke (EXPRESS) study was a prospective comparative study that found that same-day assessment and initiation of standard risk-modification measures following TIA decreased the 90-day risk of stroke from 10.3 to 2.1% (relative risk reduction of 80%) compared with delayed (or routine) out-patient evaluation and treatment [33].

Based on these data, it is clear that every effort should be made to identify TIA patients quickly and correctly, stratify stroke risk based on a thorough neurovascular evaluation and then initiate appropriate same-day treatments. Although hospitalization for all TIA patients may not be cost effective or warranted, triaging TIA patients based on diagnostic certainty followed by risk stratification may be one logical approach. Those whose risk scores are deemed to indicate a low risk of stroke (i.e., 2-day stroke risk < 5%) could be seen in specialized clinics with the capacity to start preventative treatments in the office, while those with high risk for stroke (i.e., 2-day stroke risk > 5%) should be admitted for observation and in-hospital initiation of preventative measures. However, evidence-based data regarding appropriate allocation of healthcare resources for TIA patients is sorely lacking; it is hoped that future studies will inform us of the most cost-effective and beneficial approach.

Conclusion

The importance of accurate TIA diagnosis and urgent evaluation and treatment on stroke prevention cannot be overstated. Due to its transient nature and reliance on historical features, diagnosis of TIA remains difficult. Given the current understanding of the dynamic instability following true TIA, routine or deferred out-patient evaluation can no longer be considered safe medical practice. Therefore, we advise obtaining expert neurologic consultation in all patients presenting with transient neurologic episodes, admitting high-risk patients for same-day evaluation to exclude conditions such as carotid stenosis and atrial fibrillation and initiate proven and effective therapies immediately.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.