Applicazione dell'analisi della covarianza (ancova) e della regressione logistica nella valutazione dell'immunogenicita dei vaccini antinfluenzali

The present paper describes a general method for statistical treatment of parameters to assess the antibody response to influenza vaccination. Antibody production was determined in pre- and post-vaccination sera by haemagglutination inhibition test. The results were evaluated as the geometric mean of post-vaccination titre or the mean fold increase after correction for pre-vaccination titres by analysis of covariance (ANCOVA) and as response rate or protection rate after correction for pre-vaccination titres by logistic regression. These procedures allow for the correction of the linear relationship between pre and post vaccination titres and to increase precision in determining the difference among treatment groups. Immunogenicity of trivalent sub-unit, split and whole influenza vaccines, recommended for the 1990/91 and 1991/92 winter seasons and licensed in Italy, were analysed as an example.