Objectives: The National Institute of Health and Clinical Excellence and The Royal College of Physicians recommend transferring thrombolysed patients with stroke to a hyperacute stroke unit (HASU) within 4 h from hospital arrival (TArrival-HASU), but there is paucity of evidence to support this cut-off. We assessed if a shorter interval within this target threshold conferred a significant improvement in patient mortality.
Design: We conducted a retrospective analysis of prospectively collected data from the Sentinel Stroke National Audit Programme.
Setting: Four major UK hyperacute stroke centres between 2014 and 2016.
Participants: A total of 183 men (median age = 75 years, IQR = 66-83) and 169 women (median age = 81 years, IQR = 72.5-88) admitted with acute ischaemic stroke.
Main outcome measures: We evaluated TArrival-HASU in relation to inpatient mortality, adjusted for age, sex, co-morbidities, stroke severity, time between procedures, time and day on arrival.
Results: There were 51 (14.5%) inpatient deaths. On ROC analysis, the AUC (area under the curve) was 61.1% (52.9-69.4%, p = 0.01) and the cut-off of TArrival-HASU where sensitivity equalled specificity was 2 h/15 min (intermediate range = 30 min to 3 h/15 min) for predicting mortality. On logistic regression, compared with the fastest TArrival-HASU group within 2 h/15 min, the slowest TArrival-HASU group beyond upper limit of intermediate range (≥ 3 h/15 min) had an increased risk of mortality: 5.6% vs. 19.6%, adjusted OR = 5.6 (95%CI:1.5-20.6, p = 0.010).
Conclusions: We propose three new alarm time zones (A1, A2 and A3) to improve stroke survival: "A1 Zone" (TArrival-HASU < 2 h/15 min) indicates that a desirable target, "A2 Zone" (TArrival-HASU = 2 h/15 min to 3 h/15 min), indicates increasing risk and should not delay any further, and "A3 Zone" (TArrival-HASU ≥ 3 h/15 min) indicates high risk and should be avoided.
Keywords: Evidence-based medicine; Ischaemic stroke; ROC analysis; Stroke units; Thrombolysis; tPA.