Diagnostic delays in paediatric stroke

Background: Stroke is a major cause of mortality in children. Conditions that mimic stroke also cause severe morbidity and require prompt diagnosis and treatment. We have investigated the time to diagnosis in a cohort of children with stroke.

Methods: A population-based cohort of children with stroke was prospectively identified in the south of England. Case notes, electronic hospital admission databases and radiology records were reviewed. Timing of symptom onset, presentation to hospital, first neuroimaging, first diagnostic neuroimaging and presenting clinical features were recorded.

Results: Ninety-six children with an arterial ischaemic stroke (AIS) and 43 with a haemorrhagic stroke (HS) were identified. The median time from symptom onset to diagnostic neuroimaging was 24.3 h in AIS and 2.9 h in HS. The initial imaging modality was CT in 68% of cases of AIS. CT was diagnostic of AIS in 66% of cases. MRI was diagnostic in 100%. If initial neuroimaging was non-diagnostic in AIS, then median time to diagnosis was 44 h. CT was diagnostic in 95% of HS cases. Presentation outside normal working hours resulted in delayed neuroimaging in AIS (13 vs 3 h, p=0.032). Diffuse neurological signs or a Glasgow Coma Scale <9 resulted in more expeditious neuroimaging in both HS and AIS.

Conclusions: The diagnosis of AIS in children is delayed at every stage of the pathway but most profoundly when the first neuroimaging is CT scanning, which is non-diagnostic. MRI should be the initial imaging modality of choice in any suspected case of childhood AIS.