Misdiagnosis of transient ischemic attacks in the emergency room

Abstract

Background: To determine a pattern of symptoms and/or risk factors that distinguishes transient ischemic attack (TIA) from nonischemic causes of transient neurologic attacks (NI-TNA).

Methods: We reviewed demographic, clinical, and hospital data on 100 consecutive patients with transient focal neurologic episode(s) lasting less than 24 h and in whom the initial diagnosis was TIA. After inpatient evaluation and review, final diagnoses were made by two stroke neurologists. Using stepwise multivariable logistic regression, we estimated odds ratios (OR) for independent predictors of NI-TNA. p < 0.05 was considered significant.

Results: Of the 100 patients, 40 were confirmed to have TIA and 60 NI-TNA. Compared to TIA patients, those with NI-TNA were less likely to be male and white but more likely to have a history of prior unexplained TNA, gradual symptom onset, associated nonspecific symptoms, longer symptom duration, and delayed presentation. Other variables were similar between the two groups. In a multivariable logistic regression model, gradual symptom onset (adjusted OR 6.7, p = 0.002), prior history of unexplained transient neurologic attack (adjusted OR 10.6, p = 0.031), and presence of nonspecific symptoms (adjusted OR 4.2, p = 0.008) were independent predictors of the final diagnosis of NI-TNA.

Conclusions: Distinguishing TIA from nonischemic causes is difficult in the emergency room, with 60% of suspected TIA patients having nonischemic causes on inpatient evaluation. We found 3 clinical features that may be useful in the emergency room triage of transient neurologic attacks. Further study is needed to develop tools that can accurately diagnose TIA.