Active regional surveillance for early detection of exotic/emerging pathogens of swine: A comparison of statistical methods for farm selection

Contrasting PRRSV temporal lineage patterns at the individual farm, production system, and regional levels in Ohio and neighboring states from 2017 to 2021

Porcine reproductive and respiratory virus (PRRSV), one of the most significant viruses in the swine industry, has been challenging to control due to its high mutation and recombination rates and complexity. This retrospective study aimed to describe and compare the distribution of PRRSV lineages obtained at the individual farm, production system, and regional levels. PRRSV-2 (type 2) sequences (n = 482) identified between 2017 - 2021 were provided by a regional state laboratory (Ohio Department of Agriculture, Animal Disease Diagnostic Center (ODA-ADDL)) collected from swine farms in Ohio and neighboring states, including Indiana, Michigan, Pennsylvania, and West Virginia. Additional sequences (n = 138) were provided by one collaborating swine production system. The MUSCLE algorithm on Geneious Prime® was used to align the ORF5 region of PRRSV-2 sequences along with PRRSV live attenuated vaccine strains (n = 6) and lineage anchors (n = 169). Sequenced PRRSV-2 were assigned to the most identical lineage anchors/vaccine strains. Among all sequences (n = 620), 29.8% (185/620) were ≥ 98.0% identity with the vaccine strains, where 93.5% (173/185) and 6.5% (12/185) were identical with the L5 Ingelvac PRRS® MLV and L8 Fostera® PRRS vaccine strains, respectively, and excluded from the analysis. At the regional level across five years, the top five most identified lineages included L1A, L5, L1H, L1C, and L8. Among non-vaccine sequences with production system known, L1A sequences were mostly identified (64.3% - 100.0%) in five systems, followed by L1H (0.0% - 28.6%), L1C (0.0% - 10.5%), L5 (0.0% - 14.4%), L8 (0.0% - 1.3%), and L1F (0.0% - 0.5%). Furthermore, among non-vaccine sequences with the premise identification available (n = 262), the majority of sequences from five individual farms were either classified into L1A or L5. L1A and L5 sequences coexisted in three farms, while samples submitted by one farm contained L1A, L1H, and L5 sequences. Additionally, the lineage classification results of non-vaccine sequences were associated with their restriction fragment length polymorphism (RFLP) patterns (Fisher's exact test, p < 0.05). Overall, our results show that individual farm and production system-level PRRSV-2 lineage patterns do not necessarily correspond to regional-level patterns, highlighting the influence of individual farms and systems in shaping PRRSV occurrence within those levels, and highlighting the crucial goal of within-farm and system monitoring and early detection for accurate knowledge on PRRSV-2 lineage occurrence and emergence.

A longitudinal study on PRRSV detection in swine herds with different demographics and PRRSV management strategies

Porcine reproductive and respiratory syndrome virus (PRRSV) has been one of the major health-related concerns in the swine production industry. Through its rapid transmission and mutation, the simultaneous circulation of multiple PRRSV strains can be a challenge in PRRSV diagnostic, control and surveillance. The objective of this longitudinal study was to describe the temporal detection of PRRSV in swine farms with different production types and PRRS management strategies. Tonsil scraping (n = 344) samples were collected from three breeding and two growing herds for approximately one year. In addition, processing fluids (n = 216) were obtained from piglet processing batches within the three breeding farms while pen-based oral fluids (n = 125) were collected in the two growing pig farms. Viral RNA extraction and reverse-transcription quantitative PCR (RT-qPCR) were conducted for all samples. The sample positivity threshold was set at quantification cycle (Cq) of ≤ 37. Statistical analyses were performed using generalized linear modelling and post hoc pairwise comparisons with Bonferroni adjustments using R statistical software. The results suggested a higher probability of detection in processing fluids compared to tonsil scraping specimens [odds ratio (OR) = 3.86; p = .096] in breeding farms whereas oral fluids were outperformed by tonsil scrapings (OR = 0.26; p < .01) in growing pig farms. The results described herein may lead to an improvement in PRRSV diagnostic and surveillance by selecting proper specimens.