Statistical analysis

To assess the effect of vaccination on PRRSV incidence (weaning to 6 weeks and 6–9 weeks) a generalised linear mixed‐effects model (GLMM) in Rstudio Cloud (glmer function; emmeans, Matrix and lme4 libraries) was used. Treatment (IV vaccination vs. non‐vaccination) and farm were considered fixed effects; the batch was considered as a random effect nested on the farm. Cumulative incidences from 4 to 6 and from 6 to 9 WOA and the proportion of seropositive animals at weaning were aggregated per batch and compared between vaccination groups per farm using a chi‐square test with Yates correction. Relative risk was calculated using the Koopman's likelihood‐based approximation.

Neutralisation titres (log₂ converted for normalisation of the data) were compared between PG and C groups by the ANOVA and Kruskal–Wallis tests (non‐normally distributed variables) on StatsDirect 3.2.10.

For reproductive data, a GLMM was used considering the data as binomial. In this model, pre‐weaning mortality in a litter could be influenced (fixed effects) by the treatment, the number of live piglets that were born and the parity of the sow. The interaction between the treatment and the number of piglets born alive was included since vaccination could prevent vertical transmission. Data were transformed logarithmically. Batch was used as a random effect. Since GLMM does not calculate the p‐values, these were obtained using the Wald chi‐square test (car library) and likelihood ratio tests (drop1 function).