Association of Enterovirus D68 With Acute Flaccid Myelitis, Philadelphia, Pennsylvania, USA, 2009–2018

Priyanka Uprety; Darcy Curtis; Michael Elkan; Jeffrey Fink; Ramakrishnan Rajagopalan; Chunyu Zhao; Kyle Bittinger; Stephanie Mitchell; Erlinda R. Ulloa; Sarah Hopkins; Erin H. Graf


Emerging Infectious Diseases. 2019;25(9):1676-1682. 

In This Article

Abstract and Introduction


Acute flaccid myelitis (AFM) is a polio-like disease that results in paralysis in previously healthy persons. Although the definitive cause of AFM remains unconfirmed, enterovirus D68 (EV-D68) is suspected based on 2014 data demonstrating an increase in AFM cases concomitant with an EV-D68 outbreak. We examined the prevalence in children and the molecular evolution of EV-D68 for 2009–2018 in Philadelphia, Pennsylvania, USA. We detected widespread EV-D68 circulation in 2009, rare detections in 2010 and 2011, and then biennial circulation, only in even years, during 2012–2018. Prevalence of EV-D68 significantly correlated with AFM cases during this period. Finally, whole-genome sequencing revealed early detection of the B1 clade in 2009 and continued evolution of the B3 clade from 2016 to 2018. These data reinforce the need to improve surveillance programs for nonpolio enterovirus to identify possible AFM triggers and predict disease prevalence to better prepare for future outbreaks.


A large outbreak of enterovirus D68 (EV-D68) in children in the United States caught national attention during late summer and fall of 2014. That outbreak was characterized by severe lower respiratory tract disease resulting in respiratory failure in many cases;[1] a paralytic condition occurred in a smaller subset of children.[2] Previous reports of invasive EV-D68 disease had been described across the globe,[3,4] but the magnitude of this 2014 outbreak with regard to childhood illness was unprecedented. The Centers for Disease Control and Prevention (CDC) called for national heightened awareness, particularly for rapid onset of a traumatic limb weakness along with radiologic evidence of gray matter lesions, coined acute flaccid myelitis (AFM).[2,5–7] A second period of EV-D68 circulation, again associated with severe respiratory disease, was described during late summer and fall of 2016, although 2 studies suggest lower levels of circulation during that period than in 2014.[8,9] Despite a documented spike in AFM in US children during fall of 2016 and 2018,[10] EV-D68 surveillance data from this time are limited.[11,12] In addition, little is known about the circulation of EV-D68 and association with AFM before 2014.

Although the definitive cause of AFM remains unknown, growing evidence[13,14] supports the association between EV-D68 and AFM. Epidemiologists have found that similar temporal associations between the 2014 EV-D68 outbreak and AFM cases also occurred in other countries.[14] Hixon et al. found that paralysis could be elicited in mice infected with EV-D68 strains from the 2014 outbreak.[15] As in children with AFM, these EV-D68–infected mice had a loss of motor neurons in the anterior horns of spinal cord segments corresponding to paralyzed limbs. Last, virus isolated from spinal cords of infected mice transmitted disease when injected into naive mice.[15]

Despite the mounting evidence,[13,14] the association between EV-D68 and AFM remains controversial. The detection of EV-D68 primarily in respiratory specimens with a concomitant lack of regular detection in cerebrospinal fluid (CSF)[7,16] continues to cast doubt, even though other neurotropic viruses (such as polioviruses) have similar detection patterns.[14]

What remains to be shown for EV-D68 is whether its genome has changed over time to enable increased neurotropism. Since the first description of the nonneuroinvasive EV-D68 prototype in 1962, different genetic clades associated with neuroinvasion have been identified.[15] Whole-genome sequencing (WGS) of the 2014 epidemic strains (dominated by clade B1) demonstrated a possible association between specific polymorphisms in EV-D68 and their overlap with other neurotropic and AFM-causing enteroviruses, such as poliovirus or enteroviruses D70 and A71.[16] These data suggest that the EV-D68 genome has changed over time to enable neurotropism or possibly increased virulence resulting in more widespread disease. Given that most studies have been based on partial instead of WGS, we also might be underestimating the number of clinically important genetic mutations in EV-D68 that drive increased infections with poor patient outcomes. Recognizing that longitudinal analyses of viral patterns and the molecular evolution of EV-D68 by WGS are sparse, we investigated the prevalence of EV-D68 in children in the Children's Hospital of Philadelphia (CHOP; Philadelphia, PA, USA) during 2009–2018 in relationship to AFM cases. We also investigated the phylogenetic relationships and genotypic features through WGS during this period to shed light on relevant mutations from year to year.