Transient Neurological Attack: Title and subTitle BreakA Useful Concept?

Conceptually, transient ischemic attacks (TIAs) are identical to ischemic strokes except the symptoms resolve within 24 hours. The underlying causes of TIA and stroke are identical, and recommendations for diagnostic evaluation and treatment are similar.¹ Transient ischemic attack and stroke are different, however, in another important way, distinct from symptom duration: ischemic stroke is relatively easy to diagnosis with certainty, but there is often disagreement about whether a given patient has had a TIA.

The diagnosis of TIA depends on the recollections of a patient who, by definition, is an impaired observer. Neurological deficits that constitute a possible TIA may obscure a patient's observation and recollection of the event. Even if symptoms are recalled perfectly, conditions that mimic TIA may be indistinguishable from an actual ischemic event. Syncope, seizure, migraine, peripheral vestibulopathy, and conversion can be impossible to differentiate from TIA. Consequently, even with 2 neurologists reviewing the same case, disagreement about whether a TIA occurred is common.^{2 - 3}

Accurate diagnosis of TIA is important because urgent evaluation is required for TIA, whereas most conditions that mimic TIA have been considered relatively benign. Evidence supporting the concept that a TIA is an emergency has been building over the last several years. Multiple studies have shown that the risk of stroke in the days to weeks after a TIA is high, with an average risk of 11% in the 90 days after a TIA among studies with careful monitoring from the time of TIA diagnosis.^{4 - 9} Transient ischemic attack has been variously defined in these studies, but it is almost always captured from a physician making a diagnosis in the acute setting, so the diagnosis is likely to be particularly specific, and stroke risk may be lower when the diagnosis is used more liberally.

Treating TIA emergently is not easy because treatment generally requires an emergency department evaluation and may justify hospitalization.^{1 ,10} Thus, it becomes important to get the diagnosis right: a clinician may fail to prevent a stroke if the diagnosis is missed, although clinicians may unnecessarily waste health care resources if the diagnosis of TIA is made too often in patients with other low-risk conditions, such as migraine or vasovagal syncope. Reflecting frustration in diagnostic accuracy, physicians commonly diagnose “possible TIA,” “rule out TIA,” or list TIA low in the differential diagnosis.

In this issue of JAMA, Bos and colleagues¹¹ use data from the Rotterdam study (6062 community-dwelling individuals aged 55 years and older and followed up from 1990-1993 to 2005) to examine the utility of a broad diagnostic category, transient neurological attack (TNA), based on a distinction originally suggested more than 30 years ago but rarely used today.¹² The authors define TNAs as attacks of sudden neurological symptoms that completely resolve within 24 hours, with no clear evidence for the diagnosis of migraine, epilepsy, Ménière disease, hyperventilation, cardiac syncope, hypoglycemia, or orthostatic hypotension. The definition includes TIA, which they consider focal TNA, but it also includes nonfocal TNA, a category that encompasses transient global amnesia, acute confusion, syncope without a known cause, and a variety of other conditions and neurological symptoms without clear diagnosis. Until now, nonfocal TNAs may or may not have been diagnosed as “possible TIA” but often have been treated as benign.

The idea of combining a wide array of conditions and syndromes with distinct pathophysiologies at first may seem counterproductive, but the results of the authors' analysis of TNA from the Rotterdam cohort study may argue otherwise. The authors found that patients with focal TNA (otherwise known as TIA) were at greater risk of stroke than those without TNA, which confirms prior studies of these patients.^{5 - 9} More interesting, the group with nonfocal TNA was also at increased risk of stroke and also had an increased risk of dementia, particularly of vascular dementia. A small group of patients (n = 38) with both focal and nonfocal symptoms of TNA was also at increased risk of stroke, dementia, and myocardial infarction.

The study has limitations, particularly in the diagnosis of TNA. The investigators were forced to derive the diagnosis from medical records, some of which were incomplete, and from interviews with patients designed to screen for TIA. Some TNAs were certainly missed and some events classified as TNA could probably easily have been diagnosed more specifically if diagnostic evaluations had been complete. Also, recall bias among patients with subsequent events could have increased recollection and reporting of symptoms that were classified as TNA. Many TNAs with vascular death or major stroke soon afterward were probably missed, as well. Thus, the true incidence of TNA and the risk associated with it could have been either underestimated or overestimated.

Despite these weaknesses, this is the first large-scale study of TNAs, and the findings are intriguing for several reasons. Most patients with nonfocal TNAs are currently treated as though the condition were benign. For some etiologies, such as transient global amnesia, the evidence supports this,^{13 - 14} but for most of these events, there is no consistent evaluation, no guidelines for treatment, and no information on prognosis. This study argues that, whatever is causing these events, the prognosis justifies greater attention. More needs to be done to identify the TNA patients at greatest risk, to complete evaluations, to rule out important underlying disease, and to continue to study this heterogeneous group.

Prognostic scores, such as the validated ABCD² score, use history and clinical symptoms to stratify short-term stroke risk.¹⁵ Among patients diagnosed with TIA acutely in the emergency department or clinic, the ABCD² score can identify sizable subgroups at low (<1%) and high (8%) 2-day stroke risk. Given that many of these patients are misdiagnosed and that the short-term stroke risk should be much lower among those with other etiologies for their episodes, the score is almost certainly working in part by identifying patients with true TIA. Transient ischemic attack prognostic scores might also work on the population with nonfocal TNA, but this has never been tested.

Without identifying underlying causes of nonfocal TNAs, it will be difficult to reduce the risk of subsequent stroke and vascular dementia. For patients with TIA, guidelines recommend brain imaging, carotid imaging, and an electrocardiogram for all with more detailed testing depending on findings from history and physical examination.¹ A cholesterol panel, glucose level, and hemoglobin A_1C concentration may also identify underlying risk factors for vascular disease. Given the risk of stroke in patients with nonfocal TNA, these investigations seem justified for most of them. Other evaluations might include cardiac monitoring and echocardiography. Leaving the patient unstudied and with a vague diagnosis is more difficult to justify now that the worrisome prognosis of nonfocal TNA has been demonstrated.

In terms of treatment, one could argue that the excess risk of events among patients with nonfocal TNAs may be sufficient to justify standard prevention practices recommended for patients with TIA, including an antiplatelet agent (aspirin with extended-release dipyridamole, clopidogrel, or aspirin alone), a statin, and an antihypertensive agent.¹ However, other more aggressive interventions, such as hospitalization, are probably not indicated for patients with nonfocal TNAs because the short-term risk of stroke and other adverse events is relatively low, particularly when compared with the high risk of stroke after the diagnosis of TIA.

In the end, TNA may be a term of only transient utility. This concept of TNA clarifies that more studies are needed of the large group of patients with transient neurological symptoms of sudden onset and without a clear underlying etiology. Most physicians already are appropriately concerned about patients with TIA, but the study by Bos et al extends that concern to those often diagnosed as possible TIA or left without a diagnosis. Even though TNA is likely to be only of transient utility because clinicians must quickly move to more specific diagnoses to provide appropriate treatment for patients, this entity should be considered a rally cry for more extensive evaluation or consultation in these patients, as well as for further research.