Finding PRRSV in sow herds: Family oral fluids vs. serum samples from due-to-wean pigs

The aim of this study was to compare the detection of porcine reproductive and respiratory syndrome virus (PRRSV) in due-to-wean litters in commercial swine breeding herds using family oral fluids (FOF) vs. individual piglet serum samples. FOF and piglet serum samples were collected in 199 due-to-wean litters on six farms containing 2177 piglets. All samples were individually tested for PRRSV RNA by RT-rtPCR. A litter was considered PRRSV-positive when PRRSV RNA was detected in ≥ 1 piglet serum sample or the FOF sample. Mixed effect logistic regression with farm as a random effect was used 1) to evaluate the probability of obtaining a PRRSV RNA positive FOF as a function of the proportion of viremic piglets in a litter and 2) the effect of litter size and parity on the probability that a litter would test PRRSV RNA positive in FOF. A Bayesian prevalence estimation under misclassification (BayesPEM) analysis was used to calculate the PRRSV prevalence and 95 % credible interval given the condition that all samples (FOF and serum) tested negative. In total, 34 of 199 litters (17.1 %) contained ≥ 1 viremic piglet(s), and 28 of 199 litters (14.1 %) were FOF positive. When all piglet serum samples within a litter tested negative, 1 of 165 FOF (0.6 %) tested PRRSV RNA positive. The probability of a PCR-positive FOF sample from litters with 10 %, 20 %, 30 %, 40 %, and 50 % within-litter PRRSV prevalence was 3.5 %, 35.1 %, 88.8 %, 99.2 %, and >99.9 %, respectively. The odds of a PCR-positive FOF in a first parity litter were 3.36 times (95 % CI: 2.10-5.38) that of a parity ≥ 2 litter. The odds of a positive FOF result in a litter with ≤ 11 piglets were 9.90 times (95 % CI: 4.62-21.22) that of a litter with > 11 piglets. FOF was shown to be an efficacious sample type for PRRSV detection in farrowing rooms. A risk-based approach for litter selection combined with FOF collection can be used to improve on-farm PRRSV detection with a limited sample size, compared to sampling multiple individual pigs. Finally, the BayesPEM analysis showed that PRRSV may still be present in breeding herds when all samples (serum and FOF) test PRRSV RNA negative, i.e., negative surveillance results should be interpreted with caution.