ObjectivesTo clarify diagnostic accuracy and consequences of misdiagnosis in the admission evaluation of stroke-code patients in a neurologic emergency department with less than 20-minute door-to-thrombolysis times.MethodsAccuracy of admission diagnostics was studied in an observational cohort of 1,015 stroke-code patients arriving by ambulance as candidates for recanalization therapy between May 2013 and November 2015. Immediate admission evaluation was performed by a stroke neurologist or a neurology resident with dedicated stroke training, primarily utilizing CT-based imaging.ResultsThe rate of correct admission diagnosis was 91.1% (604/663) for acute cerebral ischemia (ischemic stroke/TIA), 99.2% (117/118) for hemorrhagic stroke, and 61.5% (144/234) for stroke mimics. Of the 150 (14.8%) misdiagnosed patients, 135 (90.0%) had no acute findings on initial imaging and 100 (67.6%) presented with NIH Stroke Scale score 0 to 2. Misdiagnosis altered medical management in 70 cases, including administration of unnecessary treatments (thrombolysis n = 13, other n = 24), omission of thrombolysis (n = 5), delays to specific treatments of stroke mimics (n = 13, median 56 [31–93] hours), and delays to antiplatelet medication (n = 14, median 1 [1–2] day). Misdiagnosis extended emergency department stay (median 6.6 [4.7–10.4] vs 5.8 [3.7–9.2] hours; p = 0.001) and led to unnecessary stroke unit stay (n = 10). Detailed review revealed 8 cases (0.8%) in which misdiagnosis was possible or likely to have worsened outcomes, but no death occurred as a result of misdiagnosis.ConclusionsOur findings support the safety of highly optimized door-to-needle times, built on thorough training in a large-volume, centralized stroke service with long-standing experience. Augmented imaging and front-loaded specialist engagement are warranted to further improve rapid stroke diagnostics.