Increase in Enterovirus D68 Infections in Young Children, United Kingdom, 2006–2016

Everlyn Kamau; Heli Harvala; Soile Blomqvist; Dung Nguyen; Peter Horby; Richard Pebody; Peter Simmonds


Emerging Infectious Diseases. 2019;25(6):1200-1203. 

In This Article

The Study

For this study, we obtained serum samples collected as an approximately representative age-stratified cross section of the UK population.[10] These samples were collected in 2006, before the reports of increased number of EV-D68 cases (n = 516), and in 2016, after the 2014 EV-D68 outbreak (n = 566) (Figure 1, panel A). We used a standard microneutralization assay for serum samples (Appendix).

Figure 1.

Comparison of enterovirus D68 (EV-D68) seroprevalence in the United Kingdom in 2006 and 2016. A) Seroreactivity to EV-D68 of samples collected in 2006 and 2016 from different age categories. Results are expressed as percentage of samples displaying neutralizing antibody titers <8 or >1,024 (histogram) and geometric mean titers (red line). We performed the Kruskal-Wallace nonparametric test to evaluate differences in titer distributions between samples collected at the 2 points in each band (red text indicates p<0.05). B) Seroprevalence of neutralizing antibodies to EV-D68 in different age categories in 2006 (red) and 2016 (blue). Error bars show SEs of the proportions.

To determine the optimal strain to measure neutralizing antibody titers (Nab) assays, we compared titer of selected serum samples to the prototype Fermon (1962) strain and those of more recent EV-D68 isolates isolated in 2005 (FI_2005) and 2016 (FI_2016) (Appendix Figure 1). We identified genotypes of D for FI_2005 and B3 for FI_2016 by phylogenetic comparison of the viral protein 1 sequences. We selected samples for comparing titers to narrow down times of EV-D68 exposure: patients >40 years of age in 2006, representing serologic responses to infections acquired substantially before 2006; patients 6 months–5 years of age in 2006, representing responses to infections acquired during 2001–2006; and patients 6 months–5 years of age in 2016, representing responses to infections acquired during 2011–2016. Geometric mean titers (GMTs) of NAbs to Fermon and FI_2016 were comparable between exposure groups (Appendix Figure 2), whereas samples collected from children infected during 2001–2006 showed some evidence for proportionately higher seroreactivity to the 2005 strain. Overall, differences in GMTs were minor, and we selected the FI_2016 strain for NAb screening.

We determined seroreactivity to EV-D68 by GMT calculations for each age and year category and proportions of samples with different neutralizing antibody titers (Figure 1, panel A). We determined seroprevalence and inferred frequencies of past infection using a conservative 1:16 titer threshold (Figure 1, panel B). Frequencies and titers of EV-D68 NAbs differed substantially between the 2 collection years in young children (in the categories 0.5–1 year, 1–5 years, and 6–10 years of age). The difference narrowed in older age groups, and seroprevalence approached 100% in those >40 years of age. Seroprevalence and titer distributions were elevated in the <0.5-year age group, likely reflecting the presence of maternal antibody in these infants.

The differences in seroprevalence in the young children between 2006 and 2016 demonstrate greater infection rates in 2016. To identify the age at which this greater exposure occurred, we divided sample sets into narrower age bands and determined seroprevalence (Figure 2, panel A). For both groups, infections were acquired at a very early age, with extremely high incidences in the 0.5–2-year and 3–4-year age ranges in both sample years but a marked reduction for children >5 years of age. Annualized incidence of EV-D68 infection in the 0.5–5-year age band increased from 36 infections/1,000 population during 2001–2005 to 53 infections/1,000 population during 2012–2016, an increase of 50% (Figure 2, panel B). The increased incidence in the <5-year age group in our study would equate to >35,000 additional EV-D68 infections/year, primarily in young children 0.5–2 years of age. Incidence in older age groups was comparable or reduced, as we expected with greater rates of EV-D68 exposure and seroconversion in the younger age ranges.

Figure 2.

Comparison of incidence of enterovirus D68 (EV-D68) in the United Kingdom in 2006 and 2016. A) Estimated annual incidence of EV-D68 infection for each age group. Incidence was inferred from the difference in seroprevalence from that of the previous age band and converted into infections/year/1,000 population (by dividing the difference in prevalence by the number of years in the age band and multiplying by 1,000). Frequencies of samples with neutralizing antibody titer >16 are shown above bars. B) Change in incidence of EV-D68 infections from 2006 to 2016, expressed as additional EV-68 infections/year (y-axis scale). Figures above bars indicate the predicted positive (red) and negative (blue) change in number infections if these incidences were applied to the whole UK population, based on age-stratified population totals for 2016 obtained from Statista (