Introduction:
In-hospital stroke is associated with worse outcomes and fewer stroke interventions compared to patients with community-acquired stroke. We assessed factors associated with delays in symptom identification and stroke team alerting, and use of acute interventions for in-hospital strokes.
Methods:
The local Get With The Guidelines-Stroke and an in-hospital quality improvement database at our tertiary care hospital were screened over a 26-month period ending 10/2019, yielding 98 in-hospital strokes.
Results:
Strokes were more common on surgery services (70%), were predominantly ischemic (83%), and were moderate-to-severe (median NIHSS 16; interquartile range [IQR] 6-24). There were long delays from the time of last known normal (LKN) until stroke symptom identification (SxID) (median 5.1 hours, IQR 1.0-19.7 hours) and from SxID to stroke alert (median 2.1 hours, IQR 0.5-9.9 hours). In univariable analysis, being in an ICU, being intubated, being on a surgical service, having no lateralized weakness or neglect, and higher NIHSS were associated with delays; in multivariable analysis only intubation was associated with being above median from LKN to SxID (OR 4.3, 95% CI 1.2-16.2, p=0.03) and above median for SxID to stroke alert (OR 8.5, 95% CI 2.0-36.4). Acute stroke interventions were given to 15 patients (15%), including 3 (3%) who received IV tPA and 12 (12%) who underwent IA thrombectomy. Patients who received stroke interventions had shorter times from last normal to SxID (median 0.7 vs 8.2 hours, p=0.002) and times from SxID to stroke alert (median 0.2 vs 3.4 hours, p=0.006). Urgent vascular imaging occurred in 68/98 (69%) of patients and 23/68 (34%) had a large vessel occlusion (LVO). Of those patients with an LVO, 78% had lateralizing arm weakness or neglect on exam.
Conclusions:
There are long delays from LKN to SxID and from SxID to stroke alert in hospitalized patients leading to low rates of acute stroke treatment. Intubation was a robust risk factor for delays and protocols should be established to monitor these patients more carefully. Despite these delays, 1/3 of patients who had vascular imaging had an LVO identified that might have been eligible for intervention.