Abstract Body: Introduction: Acute infection transiently increases risk for childhood arterial ischemic stroke (AIS). We hypothesize that this paradox of a common exposure linked to a rare outcome could be explained by either (1) unusual infections—uncommon pathogens or combinations of pathogens—or (2) an unusual host response to infection. While previous studies relied solely on clinical history and targeted laboratory testing as measures of infection, we leveraged metagenomic next-generation sequencing (mNGS) as an unbiased measure of infection to test the first hypothesis.
Methods: The Vascular effects of Infection in Pediatric Stroke II (VIPS II) study is a North American and Australian prospective cohort study which enrolled children (28 days – 18 years old) with AIS at 22 sites over a 5-year period. We collected data on preceding infection via parental interview and chart abstraction at time of admission. We performed mNGS of serum and throat swabs collected within 72 hours of stroke ictus. To assess for the background spectrum of pathogens, we enrolled and performed mNGS on unmatched stroke-free well controls (WC) and ill controls (IC).
Results: Between 2017 and 2022, VIPS II enrolled 205 cases, 95 WC, and 47 IC (Table 1). Both serum and throat swab mNGS data was available for 190/205 cases, 91/95 WC, and 27/47 IC. Parents reported clinical infection in the 4 weeks prior to stroke in 47/190 (25%) cases. Chart abstraction identified hospital diagnoses of pre-stroke infection in 69/190 (35%) cases. Typical childhood viral pathogens were identified by mNGS in 26/190 (14%) cases, 9/91 WC (10%), and 9/27 (33%) IC; multiple pathogens in a single patient were rare (Table 2). Clinical infection (parental history or chart abstraction) was present in 18/26 (69%) cases with infection detected on mNGS. Overall, infection preceding or coincident with stroke was identified by parental interview, chart abstraction, or mNGS in 46% of cases.
Conclusions: Almost half of cases had evidence of infection prior to or coincident with AIS, supporting prior literature. Unbiased pathogen detection with mNGS, performed for the first time in children with AIS, detected a variety of common childhood infections in both cases and controls, suggesting that the paradoxical relationship between infection and AIS is not explained by unusual or multiple pathogens. The alternative hypothesis regarding an unusual host immune response to infection in the pathogenicity in AIS should be further explored.