Antibodies Against Egg- and Cell-Grown Influenza A(H3N2) Viruses in Adults Hospitalized During the 2017-2018 Influenza Season

Abstract

Background: Influenza vaccine effectiveness was low in 2017-2018, yet circulating influenza A(H3N2) viruses were antigenically similar to cell-grown vaccine strains. Notably, most influenza vaccines are egg propagated.

Methods: Serum specimens were collected shortly after illness onset from 15 influenza A(H3N2) virus-infected cases and 15 uninfected hospitalized adults. Geometric mean titers against egg- and cell-grown influenza A/Hong Kong/4801/2014(H3N2) virus vaccine strains and representative circulating viruses (including A/Washington/16/2017) were determined by a microneutralization (MN) assay. Independent effects of strain-specific titers on susceptibility were estimated by logistic regression.

Results: MN titers against egg-grown influenza A/Hong Kong virus were significantly higher among vaccinated individuals (173 vs 41; P = 0.01). In unadjusted models, a 2-fold increase in titers against egg-grown influenza A/Hong Kong virus was not significantly protective (29% reduction; P = .09), but a similar increase in the cell-grown influenza A/Washington virus antibody titer (3C.2a2) was protective (60% reduction; P = .02). Higher egg-grown influenza A/Hong Kong virus titers were not significantly associated with infection, when adjusted for antibody titers against influenza A/Washington virus (15% reduction; P = .61). A 54% reduction in the odds of infection was observed with a 2-fold increase in titer against influenza A/Washington virus (P = not significant), adjusted for the titer against egg-grown influenza A/Hong Kong virus titer.

Conclusion: Individuals vaccinated in 2017-2018 had high antibody titers against the egg-adapted vaccine strain and lower titers against circulating viruses. Titers against circulating but not egg-adapted strains were correlated with protection.

Keywords: Influenza A; antibodies; egg adaptation; vaccine effectiveness.

Figure 1.

Consolidation Standard of Reporting Trials diagram showing enrollment and sample availability for 168 individuals enrolled in the Hospitalized Adult Influenza Vaccine Effectiveness Network study at the time of testing.

Figure 2.

Maximum likelihood trees of hemagglutinin (HA) sequences, by clade. Branches of maximum likelihood trees of HA sequences are color-coded by clade (3C.2a1–3C.2a4 and 3C.3a; see notation near tips). The outgroup is influenza A/Texas/50/2012 virus. Samples from the Hospitalized Adult Influenza Vaccine Effectiveness Network cohort begin with “MH”; see clade 3C.2a2 (n = 19) and 4 additional sequences (black arrows). Clade assignments are based on locations of the reference sequences in nextstrain.org. Bootstrap values ≥ 70 are shown.

Figure 3.

Differences in influenza A(H3N2) virus hemagglutinins (HAs) between egg-propagated influenza A/Hong Kong/4801/2014 virus and Madin-Darby canine kidney (MDCK)–SIAT1 cell–propagated influenza A/Washington/16/2017 virus. Amino acid changes highlighted in red on the HA structure are those associated with egg adaptation in egg-grown influenza A/Hong Kong/4801/2014 virus. Potential antigenic sites are indicated in parenthesis.

Figure 4.

Estimated percentage reduction in the odds of influenza virus infection, by increasing microneutralization titers against egg-propagated influenza A/Hong Kong/4801/2014(H3N2) virus and influenza A/Washington/16/2017(H3N2) virus. A and B, Unadjusted estimates for each virus target . C and D, Estimates adjusted for the titer against the other virus. Dotted lines indicate 95% confidence intervals.