Individual risk factors for Post-weaning Multisystemic Wasting Syndrome (PMWS) in pigs: A hierarchical Bayesian survival analysis

Risk factors for Post-weaning Multisystemic Wasting Syndrome (PMWS) at the pig level were identified using data from a longitudinal study in seven PMWS-affected farms in France. In each farm, a representative sample of 120 pigs (180 in one farm) was randomly selected after farrowing and followed from birth to slaughter. Individual information included serological status for Porcine Circovirus type 2 (PCV-2), Porcine Reproductive and Respiratory Syndrome (PRRS) virus, and Porcine Parvovirus (PPV), individual weight, rearing conditions, and clinical observations recorded at 7,13,16 and 21 weeks of age and at slaughter. Two different Bayesian frailty models were used to identify variables related to time-to-PMWS: (i) a logistic-survival model and (ii) an accelerated failure time model (different survival time distributions) both with two levels of clustering (litter and farm). Similar results in terms of variables related to time-to-PMWS were obtained with both models. However, information provided by the different approaches were complementary. Piglets were more likely to exhibit PMWS after early infection by PCV-2 (i.e. before 7 weeks old) and if they were weaned early (before 21 days). Piglets born to PCV-2 seronegative sows and/or to sows with neck injuries due to poorly performed injections were also more at risk. With the accelerated failure time model, time ratios were obtained giving an estimation of the expected survival time (increased or decreased) after exposure to the factor. The logistic-survival model showed that the majority of the risk factors were mostly related to the odds of PMWS whereas the PCV-2 passive immunity derived from the dam also tended to postpone PMWS appearance later. (C) 2009 Elsevier B.V. All rights reserved.