Enterovirus D68 in Hospitalized Children: Sequence Variation, Viral Loads and Clinical Outcomes

Abstract

Background: An outbreak of enterovirus D68 (EV-D68) caused severe respiratory illness in 2014. The disease spectrum of EV-D68 infections in children with underlying medical conditions other than asthma, the role of EV-D68 loads on clinical illness, and the variation of EV-D68 strains within the same institution over time have not been described. We sought to define the association between EV-D68 loads and sequence variation, and the clinical characteristic in hospitalized children at our institution from 2011 to 2014.

Methods: May through November 2014, and August to September 2011 to 2013, a convenience sample of nasopharyngeal specimens from children with rhinovirus (RV)/EV respiratory infections were tested for EV-D68 by RT-PCR. Clinical data were compared between children with RV/EV-non-EV-D68 and EV-D68 infections, and among children with EV-D68 infections categorized as healthy, asthmatics, and chronic medical conditions. EV-D68 loads were analyzed in relation to disease severity parameters and sequence variability characterized over time.

Results: In 2014, 44% (192/438) of samples tested positive for EV-D68 vs. 10% (13/130) in 2011-13 (p<0.0001). PICU admissions (p<0.0001) and non-invasive ventilation (p<0.0001) were more common in children with EV-D68 vs. RV/EV-non-EV-D68 infections. Asthmatic EV-D68+ children, required supplemental oxygen administration (p = 0.03) and PICU admissions (p <0.001) more frequently than healthy children or those with chronic medical conditions; however oxygen duration (p<0.0001), and both PICU and total hospital stay (p<0.01) were greater in children with underlying medical conditions, irrespective of viral burden. By phylogenetic analysis, the 2014 EV-D68 strains clustered into a new sublineage within clade B.

Conclusions: This is one of the largest pediatric cohorts described from the EV-D68 outbreak. Irrespective of viral loads, EV-D68 was associated with high morbidity in children with asthma and co-morbidities. While EV-D68 circulated before 2014, the outbreak isolates clustered differently than those from prior years.

Fig 1. Sample and patient selection for the EV-D68 outbreak in 2014.

From May to November 2014 3,540 samples underwent viral testing at NCH Department of Pathology. Of those 41% tested positive for rhinovirus/enterovirus (RV/EV) by a single or multiplex PCR assay. A convenience sample of all RV/EV positive specimens was selected randomly based on availability, integrity and amount of specimen. Of this convenience sampling 44% were positive for EV-D68.

Fig 2. Number of samples tested for respiratory viruses per week in 2014.

The y-axis represents the number of samples per week. The x-axis represents the number of samples that underwent viral testing and those positive for rhinovirus/enterovirus per week from May 24 to November 29 2014. The stacked bars represent total number of respiratory samples tested for any respiratory virus. The light gray section of the bar represents the number of samples that tested negative for RV/EV while the dark gray section indicate the number of samples positive for RV/EV. The colored lines represent the convenience sampling, in blue the RV/EV-non-EV-D68 positive samples and in red the positive EV-D68 samples.

Fig 3. Neighbor-joining phylogenetic tree of EV-D68 strains.

(A) Phylogenetic tree was constructed using partial viral protein 1 gene sequences of the type strain, Fermon strain (pink circle). Sequences of EVD68 strains from the 2014 outbreak in the US and other geographic areas are depicted in blue triangles and, while sequences from Nationwide Children’s Hospital (NCH) are depicted in red diamonds (2014), purple triangles (2012) and purple squares (2011). Bootstrap values > 75% are shown. The scale bar represents genetic changes in base substitutions per site. (B) Subtree of clade B. the legend adjacent to each node reflects the accession number, the location and the year of isolation for each isolate.