Stroke Diagnosis in the Pediatric Emergency Department

See related article, p 1198. Diagnosis of pediatric stroke is often delayed, which has ramifications for the eligibility of children for hyperacute treatments such as thrombolysis and endovascular recanalization. Delayed pediatric stroke diagnosis occurs across developed countries, and delays occur at multiple stages of the prehospital, emergency room, and hospital course. In a Canadian cohort, median interval from symptom onset to diagnosis of arterial ischemic stroke (AIS) was 22.7 hours. Median delay in diagnosis was 29 hours in those with out-of-hospital strokes and 11.6 hours in those with in-hospital strokes. Interestingly, among those with out-of-hospital strokes, median time from symptom onset to hospital arrival was 1.7 hours, which indicates that most of the delay in diagnosis is on the part of medical staff. In fact, the median time from symptom onset to neuroimaging was 8.5 hours in the out-of-hospital group and 10.5 hours in the hospitalized group even though initial assessment was performed almost immediately in the hospitalized children. Lower Pediatric National Institutes of Health (NIH) Stroke Scale score, lack of seizure, and nonabrupt symptom onset were among predictors of prehospital delays. Lower Pediatric NIH Stroke Scale score also predicted delayed diagnosis in the hospitalized children. Merely 10% were diagnosed within 3 hours and 20% within 6 hours, precluding most from consideration for hyperacute therapies. In a population-based cohort from the United Kingdom, median time from symptom onset to diagnostic neuroimaging among children with AIS was 24.3 hours. A risk factor for delayed diagnosis for ischemic stroke was a normal head computed tomography scan, in which case median time to diagnosis was 44 hours. The Canadian study included children presenting from the early 1990s to the mid 2000s, but among children presenting between 2009 and 2013 in the United States, the median time from emergency department (ED) arrival to magnetic resonance imaging (MRI) was 17 hours, indicating that delayed diagnosis continues to be a major problem and barrier to timely therapy initiation. Just like physicians, parents often do not consider that a child might be having a stroke. Among 25 parents of children with confirmed AIS, despite abrupt symptom onset in 23 children, fewer than half of parents considered the possibility of stroke. Although nearly 85% of parents thought that presenting symptoms indicated a serious problem, only 36% called an ambulance. When examining elements that influence time to pediatric stroke diagnosis, understanding factors in the ED that may lead to delays is important. In this issue of Stroke, Mackay et al report on the accuracy and reliability of pediatric emergency physicians’ diagnosis of stroke among children presenting with focal neurological deficits that persisted at ED presentation. Of 287 children with 301 presentations that were evaluated in a tertiary care hospital ED, 21 had strokes (7% of presentations). Of 20 children in whom stroke was suspected, 13 (65%) had a stroke. Eight children who were ultimately diagnosed with stroke were thought to have other causes for their presentations. The sensitivity of stroke diagnosis by the ED physician was 62% with a specificity of 98%. The initial physician impression–final diagnosis κ for stroke was 0.61, which indicates substantial agreement. The physician impression–final diagnosis agreement ranged greatly for other causes of focal neurological deficits from −0.01 and 0.0 (poor agreement) for demyelination and cerebellitis to 0.96 (almost perfect agreement) for Bell’s Palsy. Although the physician impression may have been recorded before neuroimaging was performed, physician decisions about whether to obtain neuroimaging as well as type and timing of neuroimaging are likely informed by initial diagnostic considerations. The authors note that future studies should investigate the diagnostic impressions at various time points in the child’s course, that is, on presentation and then again after various investigations. The authors also compare the sensitivity of emergency physician stroke diagnosis at their center (62%) to that reported in adult studies, which can be as high as 90%. Younger adult patients are more likely to be misdiagnosed, especially if the strokes are located in the posterior circulation. In children, over 30% of strokes involve the posterior circulation, and this may contribute to the frequency of stroke misdiagnosis. Furthermore, stroke mimics are more common in the young. In 1 adult study, 21% of those younger than 50 years admitted to a stroke unit had other diagnoses. By comparison, only 3% of those older than 50 years admitted to the unit had stroke mimics. In another study of final diagnoses among 124 stroke alerts in a pediatric ED, 76% had stroke mimics. However, in a different tertiary care hospital cohort, only 21% of children evaluated by the stroke team for acute presentations concerning for stroke had mimics. Neuroimaging was required to differentiate strokes from mimics. The present study underscores the high frequency of stroke mimics in children, which makes the job of emergency physician evaluators even more difficult. Although the overall specificity of ED physician suspicion of stroke was high, the sensitivity was only moderate to good. Thus, clinical screening Stroke Diagnosis in the Pediatric Emergency Department An Ongoing Challenge