Effects of increasing IV tPA-treated stroke mimic rates at CT-based centers on clinical outcomes.

OBJECTIVE: To determine to what degree stroke mimics skew clinical outcomes and the potential effects of incorrect stroke diagnosis. METHODS: This retrospective analysis of data from 2005 to 2014 included IV tissue plasminogen activator (tPA)-treated adults with clinical suspicion for acute ischemic stroke who were transferred or admitted directly to our 2 hub hospitals. Primary outcome measures compared CT-based spoke hospitals' and MRI-based hub hospitals' mimic rates, hemorrhagic transformation, follow-up modified Rankin Scale (mRS), and discharge disposition. Secondary outcomes were compared over time. RESULTS: Of the 725 thrombolysis-treated patients, 29% were at spoke hospitals and 71% at hubs. Spoke hospital patients differed from hubs by age (mean 62 ± 15 vs 72 ± 15 years, p < 0.0001), risk factors (atrial fibrillation, 17% vs 32%, p < 0.0001; alcohol consumption, 9% vs 4%, p = 0.007; smoking, 23% vs 13%, p = 0.001), and mimics (16% vs 0.6%, p < 0.0001). Inclusion of mimics resulted in better outcomes for spokes vs hubs by mRS ≤1 (40% vs 27%, p = 0.002), parenchymal hematoma type 2 (3% vs 7%, p = 0.037), and discharge home (47% vs 37%, p = 0.01). Excluding mimics, there were no significant differences. Comparing epochs, spoke stroke mimic rate doubled (9%-20%, p = 0.03); hub rate was unchanged (0%-1%, p = 0.175). CONCLUSIONS: Thrombolysis of stroke mimics is increasing at our CT-based spoke hospitals and not at our MRI-based hub hospitals. Caution should be used in interpreting clinical outcomes based on large stroke databases when stroke diagnosis at discharge is unclear. Inadvertent reporting of treated stroke mimics as strokes will artificially elevate overall favorable clinical outcomes with additional downstream costs to patients and the health care system.