Rational clinical approach to transient ischaemic attack, stroke and associated mimics: A retrospective cohort study

Abstract

Objective: The present study was performed to identify the individual clinical features and risk factors most strongly associated with the diagnosis of transient neurological symptoms with a cerebrovascular cause (transient ischaemic attack (TIA) or stroke), as compared to common TIA mimics (including retinal ischaemia, migraine and seizure). Methods: In a retrospective cohort study, consecutive patients presenting with transient neurological symptoms to TIA clinic in Royal Adelaide Hospital were included. Clinical features and risk factors were recorded in a standardised form, categorised into subgroups, and analysed using descriptive statistics and diagnostic performance indicators, such as sensitivity, specificity and likelihood ratios. Results: For 218/1273 individuals diagnosed with stroke, the three features with the highest positive likelihood ratio were the presence of diffusion weighted imaging positive lesion on magnetic resonance imaging (23.66, 95% confidence interval (CI) 14.35–51.08), extracranial carotid atherosclerosis (3.98, 95% CI 1.19–6.87) and a history of peripheral vascular disease (3.33, 95% CI 1.64–6.27). For TIA, the three features with the highest positive likelihood ratio were extracranial carotid atherosclerosis (3.98, 95% CI 1.19–8.27), presence of atrial fibrillation (2.43, 95% CI 1.54–3.46) and pre-existing anticoagulant therapy (2.39, 95% CI 1.61–3.29). For stroke and TIA, the respective features with the lowest negative likelihood ratios were limb weakness (0.71, 95% CI 0.65–0.77) and hypertension (0.66, 95% CI 0.49–0.84). Conclusions: The present study demonstrated that specific clinical features and risk factors were associated with the final diagnosis at TIA clinic. These clinical features may assist with diagnosis of TIA in centres without access to a vascular neurologist.