-
Views
-
Cite
Cite
Guy Freeman, Benjamin J. Cowling, Serological Responses Following Influenza A(H7N9) Virus Infection, The Journal of Infectious Diseases, Volume 209, Issue 12, 15 June 2014, Pages 2018–2019, https://doi.org/10.1093/infdis/jiu021
Close - Share Icon Share
Extract
To the Editor—We commend Yang et al [1] for their important work using serology to draw new epidemiological conclusions about the recent epidemic of the novel A(H7N9) influenza virus. By determining the levels of antibody titers against this virus among poultry workers and the general population of Zhejiang province, in which a relatively high number of confirmed infections of A(H7N9) virus occurred, they found that the former have higher hemagglutination inhibition (HI) titers against the A(H7N9) influenza virus, on average, compared to the latter. This is important evidence that subclinical human infections of A(H7N9) virus have occurred among poultry workers, and further supports the role of poultry in the A(H7N9) epidemic in China. Furthermore, finding low antibody titers among the general population confirms the absence of a major “silent” A(H7N9) epidemic.
The authors also reported that postinfection antibody titers appeared to be lower among the hospitalized A(H7N9) patients who died compared to those who survived, and speculated that therefore “the presence of antibodies may improve clinical outcome in infected patients.” However, using the data kindly provided by the authors in their Supplementary Table 1B, we were not able to find evidence that survival was associated with higher antibody titers among the cohort of patients in this study, after adjustment for the time from onset of symptoms to collection of sera. We fitted a nonlinear model using the nls function of R version 3.0.1, with the logarithm of the HI titer as the outcome variable and survival status and a sigmoidal function of time since symptom onset as the explanatory variables. The fitted model is shown against the raw data in Figure 1. There is no statistically significant difference between antibody titers for the fatal vs nonfatal cases (P = .49). After including age in the regression model as a confounder, the association between antibody titer and outcome remained nonsignificant (P = .57).
