Association of outcomes in acute flaccid myelitis with identification of enterovirus at presentation: a Canadian, nationwide, longitudinal study.

Background: Acute flaccid myelitis (AFM) is characterised by rapid onset of limb weakness with spinal cord grey-matter abnormalities on MRI scan. We aimed to assess whether detection of enterovirus in respiratory or other specimens can help predict prognosis in children with AFM.

Methods: In this nationwide, longitudinal study, we evaluated the significance of detection of enterovirus in any sample in predicting outcomes in a cohort of Canadian children younger than 18 years presenting with AFM to tertiary paediatric hospitals in Canada in 2014 and 2018. All patients fulfilled the 2015 US Centers for Disease Control and Prevention case definition for definite AFM or probable AFM. Clinical data, laboratory findings, treatment, and neuroimaging results were collected (follow up period up to 5 years). We assessed neurological function and motor outcomes using Kurtzke's Expanded Disability Status Scale (EDSS) and a Weakest Limb Score.

Findings: 58 children with AFM (median age 5·1 years, IQR 3·8-8·3) were identified across five of Canada's ten provinces and three territories. 25 (43%) children had enterovirus detected in at least one specimen: 16 (64%) with EV-D68, two (8%) with EV-A71, two (8%) with coxsackievirus, 10 (40%) with untyped enterovirus. Children who were enterovirus positive were more likely than those that were negative to have had quadriparesis (12 [48%] of 25 vs four [13%] of 30; p=0·028), bulbar weakness (11 [44%] of 25 vs two [7%] of 30; p=0·028), bowel or bladder dysfunction (14 [56%] of 25 vs seven [23%] of 30; p=0·040), cardiovascular instability (nine [36%] of 25 vs one [3%] of 30; p=0·028), and were more likely to require intensive care unit admission (13 [52%] of 25 vs 5 [17%] of 30; p=0·028). On MRI, most children who were enterovirus positive showed brainstem pontine lesions (14 [61%] of 23), while other MRI parameters did not correlate with enterovirus status. Median EDSS of enterovirus positive (EV+) and enterovirus negative (EV-) groups was significantly different at all timepoints: baseline (EDSS 8·5, IQR 4·1-9·5 vs EDSS 4·0, IQR 3·0-6·0; p=0·0067), 3 months (EDSS 4·0, IQR 3·0-7·4 vs EDSS 3·0, IQR 1·5-4·3; p=0·0067), 6 months (EDSS 3·5, IQR 3·0-7·0 vs EDSS 3·0, IQR 1·0-4·0; p=0·029), and 12 months (EDSS 3·0, IQR 3·0-6·9 vs EDSS 2·5 IQR 0·3-3·0; p=0·0067). Kaplan-Meier survival analysis of a subgroup of patients showed significantly poorer motor recovery among children who tested positive for enterovirus than for those who tested negative (p=0·037).

Interpretation: Detection of enterovirus in specimens from non-sterile sites at presentation correlated with more severe acute motor weakness, worse overall outcomes and poorer trajectory for motor recovery. These results have implications for rehabilitation planning as well as counselling of families of children with these disorders. The findings of this study support the need for early testing for enterovirus in non-CNS sites in all cases of AFM.

Funding: None.