To study the frequency of stroke mimics and analyze the predictive value of the 2CAN score for accurate diagnosis of stroke among in-patient stroke alerts.

The stroke alert (SA) process has standardized the diagnosis and management of acute stroke (AS) with prompt delivery of time-sensitive treatment and intervention.  For in-patients, the process may be suboptimal due to the high incidence of stroke mimics (SM), both medical or neurological  leading to delayed diagnosis and care. The 2CAN-score is a novel clinical tool with good sensitivity (92%) for identifying stroke among in-patients. We examined the frequency of SM and the predictive value of 2CAN score for diagnosis of AS among an in-patient cohort for whom SA was activated.

Consecutive in-patient SAs from August 2020-March 2021 at our tertiary care stroke center were reviewed. Demographic, historic and examination findings were used to calculate 2CAN- score (clinical deficits: speech abnormality, facial droop, and asymmetric arm weakness, recent cardiac procedure, history of atrial fibrillation, new admission < 24 hours), final diagnosis per brain CT/MRI were retrospectively collected. The positive predictive value (PPV) and negative predictive value (NPV) of 2CAN score > 2 for diagnosis of AS were estimated.

About two-thirds of in-patient SA occurred for SM. 2CAN score can be a sensitive tool to triage the hospitalized patients deemed low-risk for stroke for whom SA activation may not be warranted. Implementation of this tool can help optimize the process allowing appropriate diagnosis and management of conditions presenting as stroke mimics.