Molecular and Clinical Comparison of Enterovirus D68 Outbreaks Among Hospitalized Children, Ohio, USA, 2014 and 2018

Huanyu Wang; Alejandro Diaz; Katherine Moyer; Maria Mele-Casas; Maria Fatima Ara-Montojo; Isabel Torrus; Karen McCoy; Asuncion Mejias; Amy L. Leber


Emerging Infectious Diseases. 2019;25(11):2055-2063. 

In This Article


Sample Selection and Clinical Characteristics of EV-D68–Infected Patients during the 2018 Outbreak

During the 2018 outbreak, of 3,633 samples tested by the FilmArray panel 1,987 (55%) were positive for RV/EV. We further evaluated 1,028 samples, of which 401 (39%) tested positive for EV-D68. This number compares with 213 (46%) EV-D68–positive samples of 459 in the 2014 outbreak (Figure 1).

Of the 401 patient samples that tested positive for EV-D68 in 2018, we excluded 124 (122 because patients were evaluated in the outpatient setting or clinical data were not available and 2 from children with nosocomial EV-D68 infection), leaving a total of 277 patients hospitalized with EV-D68 infection. Of those, 67 (24%) children were previously healthy, 146 (53%) had preexisting asthma or a history of wheezing, and 64 (23%) had another underlying chronic medical condition (Table 1). Children with a history of asthma or wheezing were older (median age 4.1 years) than children who had complex medical conditions (median 2.5 years) or were previously healthy (median 1.4 years; p<0.01); findings did not differ by sex or race. Most children had respiratory symptoms (94%–100%) with or without fever, followed by gastrointestinal manifestations (27%–32%). Eight (2.9%) children had neurologic manifestations, 2 AFM, and 1 opsoclonus/myoclonus syndrome (OMS).

Overall, children with asthma required pediatric intensive care unit (PICU) admission more frequently (63%) than did previously healthy children (37%) or children with chronic medical conditions (29%; p<0.0001); however, duration of hospitalization was longer for children with underlying conditions. The EV-D68 semiquantitative viral load (cycle threshold) ranged from an average of 25.1 to 26.1 and did not differ significantly between groups.

Clinical Manifestations During the 2014 and 2018 EV-D68 Outbreaks and the 2016 Low-activity Period

We compared the demographic and clinical characteristics of children from the 2014 and 2018 EV-D68 outbreaks (Table 2). Overall, children with EV-D68 infection identified in 2018 were significantly younger and more often of white race; we found no differences in sex or presence of asthma or other chronic medical conditions. During the 2018 outbreak, children with EV-D68 infection more commonly had gastrointestinal symptoms than during the 2014 outbreak (28% in 2018 vs. 12% in 2014; p<0.001); symptoms included emesis, abdominal pain, and diarrhea. On the other hand, children with EV-D68 infection identified during the 2014 outbreak had respiratory symptoms and skin rashes more frequently than did children during the 2018 outbreak. The proportion of children who required PICU admission was lower in 2018 (49%) than in 2014 (68%; p<0.0001), and duration of hospitalization was shorter in 2018 (2.5 days) than in 2014 (2.8 days; p = 0.01).

Severe neurologic manifestations occurred more often during the 2018 outbreak. In the 2014 cohort, 4 (2%) patients had febrile seizures, but no other neurologic findings were documented, and no case of AFM was identified. In 2018, however, 8 (2.9%) patients had neurologic manifestations; 2 had AFM and 1 OMS. Two of these 3 children were previously healthy; 1 had underlying asthma. Four additional children sought treatment for complex febrile seizures (3 of these patients had a history of epilepsy), and 1 infant had viral meningitis. In this infant, parechovirus was identified by rRT-PCR in cerebrospinal fluid (Appendix Table 1,

During 2016, of 3,098 samples tested by the FilmArray panel, 1,293 (42%) were positive for RV/EV. Of those, 211 were further tested for EV-D68, and 14 (7%) yielded positive results. Nine of the 14 patients identified with EV-D68 infection were hospitalized; all had respiratory symptoms, and none had neurologic manifestations. Equal proportions of children had asthma or chronic medical conditions or were previously healthy (3 [33%] each).

EV-D68 Seasonality and Asthma

We also compared the proportion of samples in which we detected EV-D68 during June–October 2014, 2016, and 2018 and analyzed the seasonality of EV-D68 in relation to admissions for asthma (Figure 2). The proportion of RV/EV detected from all nasopharyngeal samples analyzed according to the standard of care was 41% in 2014 and 2016 and 55% in 2018; overall EV-D68 detection was 44% in 2014, 7% in 2016, and 39% in 2018. The proportion of EV-D68 detected among RV/EV-positive samples (Figure 2, panel A) and the number of admissions for asthma (Figure 2, panel B) were calculated weekly during the 3 periods. The duration of the 2014 outbreak was shorter (mid-July through early October, 13 weeks), and peaked the last week of August, coinciding with a disproportionate number of admissions for asthma. In 2018, the first cases of EV-D68 were identified earlier (mid-June) and ended the first week of October (total of 16 weeks), peaking the last week of July/first week of August. In parallel, the number of admissions for asthma increased during July and August but peaked during the last week of August, which was delayed in relation to EV-D68 circulation. Admissions for asthma during the same time period in 2016 were substantially lower, as was EV-D68 detection.

Figure 2.

Percentage of EV-D68 (A) and number of admissions for asthma per 1,000 hospital admissions (B) among rhinovirus/enterovirus-positive (RV/EV) samples, Nationwide Children's Hospital, Columbus, Ohio, USA, June–October 2014, 2016, and 2018. EV-D68, enterovirus D68.

Molecular Characteristics of EV-D68 2018 Strains

We sequenced 130 EV-D68–positive samples from 2018 and aligned them to both NCH strains from prior years and to EV-D68 sequences available at the National Center for Biotechnology Information website. The NCH 2018 strains were >98.5% identical to each other and demonstrated >85% sequence identity to the VP1 regions of the prototype Fermon strain (GenBank accession no. NC_038308). These 2018 strains were also 92%–94% identical to the 2014 and 2011 NCH strains previously reported.[3] The most closely related sequences to the NCH 2018 strains were those isolated in 2016 and 2015 from different geographic regions.[10,11]

We used 17 NCH strains (10 from 2018, 5 from 2014, and 2 from 2011) for further genetic characterization. Phylogenetic analyses followed by bootstrap analyses indicated that all NCH strains identified during the 2018 outbreak clustered into a new sublineage within major clade B, differently from the 2011 and 2014 NCH strains (Figure 3). Amino acid sequence alignment for the BC, DE, and GH loops (Figure 4) showed that the NCH 2018 strains displayed a unique amino acid signature, and all contained the amino acid residue (218T) that characterizes the EV-D68 clade B3.[10]

Figure 3.

Phylogenetic analysis of EV-D68 from samples from children at Nationwide Children's Hospital, Columbus, Ohio, USA, 2011, 2014, and 2018. Phylogenetic tree was constructed using partial viral protein 1 gene sequences. Scale bar indicates changes in base substitutions per site. EV-D68, enterovirus D68.

Figure 4.

Amino acid analysis of 17 EV-D68 strains from samples from patients at NCH, Columbus, Ohio, USA, 2011, 2014, and 2018. EV-D68 strains representing subclades B1, B2, and B3 were aligned. Black boxes indicate amino acids included in viral protein 1 motifs corresponding to protein loops; colored boxes indicate strains corresponding to each subclade (green, clade B1; blue, clade B2; red, clade B3). GenBank accession numbers are given in parentheses. EV-D68, enterovirus D68; NCH, Nationwide Children's Hospital.