Background: Influenza vaccination strategies aim at protecting high-risk population from severe outcomes. Estimating the effectiveness of seasonal vaccines against influenza related hospitalisation is important to guide these strategies. Large sample size is needed to have precise estimate of influenza vaccine effectiveness (IVE) against severe outcomes. We assessed the feasibility of measuring seasonal IVE against hospitalisation with laboratory confirmed influenza through a network of 21 hospitals in the European Union.
Methods: We conducted a multicentre study in France (seven hospitals), Italy (one hospital), and Navarra (four hospitals) and Valencia (nine hospitals) regions in Spain. All ≥18 years hospitalised patients presenting an influenza-like illness within seven days were swabbed. Cases were patients RT-PCR positive for influenza A (H3N2); controls were patients negative for any influenza virus. Using logistic regression with study site as a fixed effect we calculated IVE adjusted for potential confounders. We restricted the analyses to those swabbed within four days.
Results: We included, 375 A(H3N2) cases and 770 controls. The overall adjusted IVE was 24.9% (95%CI-1.8;44.6). Among the target group for vaccination (N = 1058) the adjusted IVE was 28.8% (95%CI:2.8;47.9); it was respectively 36.8% (95%CI:-48.8; 73.1), 42.6% (95%CI:-16.5;71.7), 17.8%(95%CI:-40.8; 52.1) and 37.5% (95%CI:-22.8;68.2) in the age groups 18-64, 65-74, 75-84 and more than 84 years.
Discussion: Estimation of IVE based on the pooling of data obtained through a European network of hospitals was feasible. Our results suggest a low IVE against hospitalised confirmed influenza in 2011-12. The low IVE may be explained by a poor immune response in the high-risk population, imperfect match between vaccine and circulating strain or waning immunity due to a late season. Increased sample size within this network would allow more precise estimates and stratification of the IVE by time since vaccination and vaccine types or brands.