The porcine reproductive and respiratory syndrome (PRRS) is one of the most economically important swine endemic disease in the United States, causing losses of about $664 million annually. One common practice adopted in the U.S. is to eliminate the virus from breeding herds after an outbreak. However, anecdotal reports demonstrating large variability within and between production systems in the time needed to achieve stability and perception for the presence of multiple PRRS virus (PRRSV) strains circulating in breeding herds were common among producers and veterinarians. The sequencing of the PRRSV open reading frame 5 (ORF5) region that represents only 4% of the whole-PRRSV genome is routinely used in the U.S. to investigate and understand the PRRSV genetic variability within and between breeding herds. The ORF5 sequence is recovered with the Sanger technique, and the output is a consensus sequence that cannot detect if more than one strain is present in the sample. On the other side, next-generation sequencing (NGS) seems to be promising by targeting the sequencing of the whole-PRRSV genome, and can detect multiple distinct strains when present in the sample. Processing fluids, a novel population-based sampling specimen, has been recently adopted as part of the surveillance strategy to detect PRRSV in newborn piglet populations at breeding herds undergoing PRRSV elimination.

The objective of this work was to assess the effect of whole-PRRSV genetic variability on time-to-low prevalence in breeding herds undergoing PRRSV elimination. A total of 20 breeding herds that faced a PRRS break and adopted measures of load-closure-exposure to eliminate the virus from the herd were retained for this study. At time of outbreak serum or tissue samples were submitted for NGS, and recovered sequences were used to compare within and between herd PRRSV genetic variability. For 16 out of 20 enrolled herds, a whole-PRRSV genome was recovered and set as farm referent strain. In 4 herds only, contigs, i.e., genome fragments, were recovered. In one herd (1 of 4), a large contig of 12,888 bp was recovered and used as farm referent strain. The mean whole-PRRSV genome nucleotide identity between herds (n=16) was at 86.14 %. Processing fluid samples were submitted for NGS after the break, around 10 weeks, and the last samples with Ct <30 before farm achieved stability. The success rate of whole-PRRSV genome (~15.1 kb) recovery in processing fluid was 5.26% (5 out of 95). However, we develop a new technique as part of this study, using contigs (i.e., portions of the genome) to characterize the genetic diversity of PRRSV in the samples. The contig-based analyses revealved at least two wild-type PRRSV strains circulating simultaneously in 18 farms and at least one vaccine-like strain in 8 of 18 farms. Farms with ≤ 2 PRRSV strains reached stability before farms ≥ 3 PRRSV strains (log-rank 0.0459). The median time to the low prevalence and interquartile range (25th to 75th percentiles) for herds having ≤2 PRRSV strains and ≥ 3 strains were 30.76 weeks (27.0, 37.0) and 42.80 weeks (26.0, 54.0), respectively. PRRSV recombination events were detected in 11 farms (11 of 17), been 9 of them between wild-type strains and 2 of them between wild-type and vaccine-like strains. The recovery of PRRSV sequences by applying NGS in processing fluid samples provided scientific evidence supporting the presence of multiple PRRSV strains circulating simultaneously in breeding herds. NGS results are not restricted to the ORF5 region and provided better insights into the PRRSV genetic variability circulating in breeding herds. Surprisingly the detection of at least two wild-type PRRSV strains in a breeding herd (18 of 20) and recombination events (11 of 17) was common. The presence of higher genetic variability in a breeding herd, i.e., ≥ 3 strains, was associated with a 12-week increase in the median time to low prevalence when compared with herds that had ≤ 2 strains detected. The novel way to visualize NGS outcomes provided better insights into the PRRSV dynamics and genetic variability present in breeding herds.

Also, funds from this proposal were leveraged with an IPPA-funded project (upon approval of both entities). A more detailed final report is currently being prepared to share with IPPA (due December, 2021). In summary, time-to-low-prevalence (TTLP) was measured, defined as reaching 8 consecutive weeks of processing fluids-negative PCR. The median (25th-75th percentile) TTLP was 36 (32 – 50) weeks. This number is approximately 9 weeks longer than a previous study that our group conducted 10 years ago. Reasons for the longer TTLP may include different viral populations (i.e., PRRSV envolving to gain persistence at the population level), a more sensitive sampling scheme (processing fluids outperform the herd sensitivity of serum-based monitoring), or a change in management practices including herd closure-associated practices, immunization practices, and bio-management activities. These farms were also monitored for time-to-baseline-productivity (TTBP), defined as number of weeks it took for farms to recover the same quantity of piglets weaned per week that the farm had prior to the PRRSV outbreak. The median and 25th-75th percentiles were 22 (16 – 26), which is also longer than that observed with a cohort of 60 herds monitored 10 years ago: 16 (8,18). The TTBP metric was calculated using statistical process control, the same methodology as the original study. Finally, the total loss (median, 25th – 75th percentile) measured as number of pigs per 1,000 sows weaned below the average prior to the outbreak was 4,144 (2,363 – 5,557), also higher than the numbers observed 10 years ago: 2,789 (1,174-4,755). The longer TTBP and higher total losses indicate opportunities for additional investigation on risk factors, including method of exposure, virus characteristics, and disease control-associated practices including herd closure and internal biosecurity.

Key Findings:

  • Next-generation sequencing applied in processing fluid provided better insights into PRRSV dynamics and genetic variability within a breeding herd
  • The presence of at least two wild-type PRRSV strains in a breeding herd was common (18 of 20)
  • The presence of higher genetic variability in a breeding herd, i.e., ≥ 3 strains, was associated with a 12-week increase in the median time to low prevalence when compared with herds that had ≤ 2 strains detected
  • PRRSV recombination events are common
  • Compared to a cohort of breeding herds investigated 10 years ago, contemporary herds had longer time to low prevalence (+9 weeks), longer time to baseline productivity (+6 weeks), and more severe losses (+1,355 pigs not weaned per thousand sows).