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

Materials and Methods

Sample Collection and Testing Algorithms

We identified children hospitalized at NCH who had EV-D68 infection during the 2018 outbreak and during a nonoutbreak period (2016) and compared their clinical manifestations and characteristics with those identified during the 2014 outbreak as previously described.[3] In brief, during June 1–October 19, 2018, we collected nasopharyngeal samples using flock swabs that went into viral transport media. All samples were obtained in accordance with standard of care for patients <21 years of age who tested positive for rhinovirus/enterovirus (RV/EV) on the FilmArray Respiratory Panel version 1.7[6] and were stored at −80°C for further testing. We retrospectively identified samples from 2016, collected during the same period, and retrieved them from −80°C storage for testing. After excluding duplicates, we selected samples on the basis of availability, amount of remnant, and integrity for EV-D68 testing using a laboratory-developed real-time reverse transcription PCR (rRT-PCR) targeting the 5' nontranslated region of the human enterovirus genome as described.[3] Because we conducted EV-D68 testing after patient encounters, results were not available to the treating physician.

The sampling selection criteria differed between the 2 outbreak periods (Figure 1). During 2014, we screened a smaller set of samples for EV-D68 and focused on hospitalized patients. In 2018, we screened samples from both outpatients and inpatients. However, for both periods the clinical analyses focused on hospitalized patients only.

Figure 1.

Sample and patient selection for investigation of EV-D68 outbreaks, Columbus, Ohio, USA. Viral testing was conducted at Nationwide Children's Hospital Department of Pathology. During May–November 2014, a total of 3,540 samples underwent viral testing, of which 41% tested positive for RV/EV by a single or multiplex PCR. Four hundred fifty-nine samples were selected randomly on the basis of availability, integrity, and amount of specimen, of which 44% were positive for EV-D68. During May–October 2018, a total of 3,633 samples were tested for RV/EV by FilmArray Respiratory Panel v1.7 (6); 1,987 (55%) were positive. Of the 1,025 convenience samples, 401 (39%) were positive for EV-D68. After samples for which clinical data were not available, for which patient age was ≥21 years, or for which EV-D68 was acquired nosocomially were excluded, 192 case-patients from the 2014 outbreak and 278 from the 2018 outbreak were included in the analyses. EV-D68, enterovirus D68; RV/EV, rhinovirus/enterovirus.

We reviewed electronic healthcare records from patients positive for EV-D68 for data collection. Patients in whom EV-D68 was identified but for whom clinical data were not available, those evaluated in the outpatient setting, those ≥21 years of age, and patients who acquired RV/EV infection during hospitalization were excluded from analyses. We compared differences in clinical characteristics and disease severity parameters among patients from the 2018 EV-D68 outbreak, those identified in 2016, and patients during the 2014 outbreak. The clinical characteristics of patients during the 2014 outbreak were previously reported.[3] The Institutional Review Board of NCH approved the study.

Admissions for Asthma

We retrieved data related to admissions for asthma during the same time periods in 2014, 2016, and 2018 from the electronic data warehouse. We included patients <21 years of age who were hospitalized with an asthma diagnosis in any NCH inpatient unit. We used the following codes from the International Classification Diseases, Ninth (ICD-9) or Tenth (ICD-10) Revision, for asthma: ICD-9, 493.*; ICD-10, J45.20–J45.998.

EV-D68 rRT-PCR Testing and Sequencing

For EV-D68 detection and quantitation, we used a laboratory-developed rRT-PCR as described previously.[3] We selected a subset of EV-D68–positive samples for partial viral protein (VP) 1 gene sequencing of an 805-bp PCR product, as described.[7] We performed cycle sequencing with BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, on the automated sequencer 3130xl Genetic Analyzer (Applied Biosystems) bidirectionally. We generated multiple sequence alignments and phylogenetic trees and compared the amino acid sequences (including BE, DC, and GH loops) of partial VP1 as described.[3,8,9]

Statistical Analysis

We conducted descriptive analyses using frequency distributions or rates and used medians and interquartile ranges to summarize the demographic data and patients' baseline characteristics. We analyzed associations between categorical variables using the Fisher exact or χ 2 test and assessed normality for continuous variables using the Shapiro-Wilk test and 2-tailed Student t tests, Mann-Whitney test, 1-way ANOVA (analysis of variance), or Kruskal-Wallis tests where used as appropriate. Two-tailed p values <0.05 were considered significant. We performed statistical analyses using GraphPad Prism 8 (