# Production Losses From an Endemic Animal Disease: Porcine Reproductive and Respiratory Syndrome (PRRS) in Selected Midwest US Sow Farms

1670
• 1Department of Agricultural and Resource Economics, University of California Davis, Davis, CA, United States
• 2Department of Veterinary Population Medicine, College of Veterinary Medicine, University of Minnesota, Saint Paul, MN, United States
• 3Center for Animal Health and Food Safety, University of Minnesota, St. Paul, MN, United States

## 3. Results

Sixteen farms fulfilled the inclusion criteria for this study, leading to a balanced dataset with 768 weekly observations distributed over 48 successive weeks. All outbreaks occurred during the second half of 2014. During the pre-outbreak period, sow inventory on different farms ranged from 2,714 to 6,009 breeding females (mean = 4,245, SD = 696) (Figure 1). Among our sample, only two farms (ID = 6 and 11) stood out as having unusually large or small sow inventories (Figure 1).

Weekly observed values indicated weaned pig production on every farm worsened sharply at or just before the PRRS outbreak, as did the seven performance indicators. That effect lasted for some weeks (Figure 2).

FIGURE 2

Figure 2. Weekly average (red lines) and observed (dots) number of weaned pigs and the seven additional performance indicators during pre- and post-outbreak period.

### 3.1. Estimation of Production and Performance Baselines

The estimated δ’s were close to zero and/or statistically insignificant (P ≥ 0.05) in the regression for pre-outbreak number of weaned pigs (WP) and for nearly all of the performance indicators, indicating pre-outbreak production stability. The exceptions occurred in the regressions for abortions (AB) and litter size (LS) (See Supplementary Material 1, Table S1). After analyzing the data, we dropped the observation for t − 1 in the regression for the number of sows aborting and the observations for weeks t − 1 and t − 2 for the number of live births per litter, after which the estimated δ’s for those performance indicators were close to zero and insignificant as well. We used the means of the fitted variables for these regressions to establish their respective baselines (Table 1).

TABLE 1

Table 1. Mean baseline estimates for weaned pigs production and performance indicators

### 3.2. Estimation of PRRS Impacts

Dummy coefficients for weeks (Tt) were statistically different than 0 (P < 0.05) indicating that the inclusion of Tt as fixed time effects in equation (2) is appropriate. Although we found no significant trend in weaned pig production during the pre-outbreak period, holding the number of sows and season constant when using equation (2), we observed a consistent decrease (14 of 16 farms) in weaned pig production relative to the baseline in the week t − 1, immediately before the outbreak was reported in week t (See Table S2 and Figure 3). The decreases ranged between 1 and 12%. As our regressions based on equation 1) showed no significant trend in weaned pig production even when week t  1 was included, we did not remove week t − 1 from the baseline period. Had we done so, the baseline would have been very slightly higher and the estimated damages from PRRS slightly greater, as discussed subsequently.

FIGURE 3

Figure 3. Means of fitted number (log) of weaned pigs during pre- and post-outbreak period (Table S2). The horizontal dotted line shows the estimated overall baseline of production (Table 1). The smooth blue line shows a 4th degree polynomial function that fits estimated means of number (log) of weaned pigs each week during the post-outbreak period (t to t + 35). Vertical lines indicate 95% confidence intervals.

We estimate that aggregate weaned pig production for the 16 farms decreased from the baseline production of 2,094 per week to 1,600 in week t + 5, when output was a full 23% lower than the baseline. Table 1. The results show that farm production decreased monotonically from t − 1 to t + 5, and then began to recover (see Figure 3 and Table S2). Output recovered moderately from t + 5 until t + 11, at which point another significant decline in production occurred to t + 17 (see Figure 3 and Table S2). Eight of the 16 farms then recovered monotonically to their baseline production levels by t + 33, but a slight drop occurred again in t + 34 and t + 35 with 15 farms producing lower than the baseline. In the aggregate, observed production approached the baseline value by the end of t + 35, when our sample ended. Estimated output appears slightly lower than the pre-outbreak level, but the difference is not statistically significant (Figure 3 and Table S2).

Similar to the production of weaned pigs, the seven performance indicators did not fully recover to their pre-outbreak means. Week-to-week comparisons revealed changes in all performance indicators, with some variation in timing and intensity. For each performance indicator, the recovery of each farm fluctuated around a rising trend estimated for all farms, and again showed a non-monotonic recovery (Figure 4). As expected, some performance indicators presented a lag with respect to the trend observed in weaned pig production. A significant increase in the number of pre-weaned pigs dead was detected at t with an average expected rise of 79 (SE = 19) deceased animals relative to the baseline, reaching a maximum increase at t + 1, with 143 expected extra losses (SE = 23) (Figure 4). While litter size did not show a significant decline at t, the expected number of live births decreased by around 1 animal between t + 1 and t + 18, reaching a maximum decline at t + 2 and t + 3 and a new deterioration at t + 14. The number of stillbirths increased between t and t + 16, reaching a maximum at t + 12, with 2 stillbirths per litter (see Table S2 and Figure 4). Although there was no immediate increase in the number of sows designated for repeated service the week of the outbreak report, by t + 6 the number of sows that were designated to repeat service increased from 11 sows in the estimated baseline to 31 sows. Likewise, the number of pigs farrowed declined after t + 1. The number of abortions significantly increased at week t − 1, doubling the number of sows that aborted prior to t − 1. The number of sows with abortions peaked in the week of the outbreak report (i.e., at t) at a level five times higher than the baseline level. Finally, sow mortality showed a significant increase in t + 1 with one more sow death than during the baseline period (see Table S2 and Figure 4). In general, the indicators confirm that a PRRS outbreak affected several production stages (e.g., pre-mated sow, pregnant sow, and piglets) for an extended period of time (see Table S2).

FIGURE 4

Figure 4. Means of fitted numbers (log) of the seven performance indicators during pre- and post-outbreak period (Table S2). The horizontal dotted lines show the estimated baselines of each indicator (Table 1). The smooth blue lines show a 4th degree polynomial function that fits estimated means number (log) for each indicator during the post-outbreak period (t to t + 35). Vertical lines indicate 95% confidence intervals.

### 3.3. Production Efficiency

Our estimates indicate that a PRRS outbreak caused a 7.4% (min = 4.1%, max = 13.4%) decrease in weaned pigs per sow year, i.e., 1.92 (min = 1.05, max = 3.18) fewer weaned pigs per breeding unit. The causes of the decrease can be seen in the performance indicators. There was a slight decrease (2.4%) in the number of sows farrowing per year, from 2.3 to 2.2 farrows per sow year. In an average sized farm of this firm (i.e., 4,245 sows, see Figure 1), the slight reduction in farrowing yielded a decline of 249 fewer farrows per year. The chances that a sow repeats service increased by 37%, while aborted fetuses increased by 26% in a year with a PRRS outbreak (Table 2).

TABLE 2

Table 2. Estimated means of sow efficiency comparing years in absence and presence of a PRRS outbreak.

### 3.4. Timing of Outbreak

The decline in weaned pigs marketed in week t − 1, although statistically insignificant, as well as changes in some performance indicators (e.g., number of sows aborting), suggest that the outbreak may have started in week t − 1, one week before it was reported. We therefore developed an alternative estimate of production losses that can be compared to the estimated loss if the outbreak is assumed to begin in week t. Eliminating t − 1 from the pre-outbreak period led to estimation of a slightly higher baseline and, as a result, to a higher estimate of PRRS losses. Nonetheless, the difference between this estimate and our primary estimate is very small. Our primary estimate (using 12 weeks as pre-outbreak period) is that PRRS reduced weaned pig production per farm by 7.4% on an annual basis, leading to a decrease in output value per sow year of $86.6, or$367,521 per farm year for an average sized farm. If instead we assume the outbreak began in t −1 (i.e., using 11 weeks as pre-outbreak period), the estimated reduction in weaned pig production was 7.6%, or $88.8 less per sow year and an average revenue loss of$376,773 among the farms studied.

## 4. Discussion

We analyzed the impact of a PRRS outbreak on weaned pig production in a set of sow farms that are part of the same swine firm in the US. We estimated the time profile of disease effects, identifying the weekly changes in output relative to a pre-outbreak baseline. We find that PRRS caused a 7.4% decline in production value measured over a one-year period. Correspondingly, PRRS reduced production by 1.92 weaned pigs per sow when adjusted to an annual basis. This decrease is substantially larger than the 1.44 decrease of weaned pigs per sow/year reported in another study (9). We note that total losses due to PRRS are likely to be greater than the revenue losses estimated in this study, as total losses must include cost increases associated with the disease, e.g., an increase in management expenses, biosecurity investments, additional feed and veterinary inputs, plus a possible decrease in the weight or in the sales price of piglets (4).

We found that weaned pig production declined in week − 1, although statistically insignificant, as did several performance indicators. The data suggest that the average PRRS outbreak in this set of farms began at least one week before it was announced. This delay may be explained, at least in part, by the inability of producers to detect PRRS until animals begin to show explicit clinical signs, as well as the additional time needed to test and confirm the disease. The lag between the outbreak of disease and the appearance of clinical signs may be longer in farms using vaccination programs, as in our sample, where clinical signs may be subtle (2631). It seems likely that some weaned pigs being shipped by these farms in week t − 1, when the disease was almost certainly present in these farms, but as yet unannounced, were infected with PRRS. The relatively slow identification of the disease means that animal movements out of infected premises must be a common source of disease spread. This is particularly important in sow farms that deliver wean pigs to different swine grower facilities each week. Reducing disease spread via movements of diseased animals might significantly reduce overall losses to PRRS (10).

The rise in abortions was the strongest signal of PRRSV activity in our data. Increased surveillance, particularly to rising abortions, may allow farms to identify PRRS more quickly. Abortions were rising in the several weeks prior to the reporting of the outbreak in some of the farms in the sample. Abortions rose significantly in t − 1 and then increased sharply in week t. The number of abortions declined rapidly and fairly monotonically following week t, with a slight uptick in weeks t + 10 to t + 13, and recovered to the baseline level by about week t + 20. Thus, to the extent that abortions are an indicator of an active virus in the sow herd, circulation of the virus appears to have ended about 20 weeks after it was reported. The uptick in weeks t + 10 to t + 13 suggests that the disease may have been infecting other susceptible cohorts of sows within the farms two to three months after the initial outbreak.

The length of PRRS outbreaks, as well as their effects over time, is highly variable. For example, one study estimated effects of an outbreak during 12 weeks post detection (32), while another indicated that production of negative piglets was reached 27 weeks post infection (21). Our results demonstrate that PRRS has a negative effect on weaned pig production for a longer time than previously estimated. In our study, the estimated means of weaned pig production remained below the baseline throughout the 35 weeks that we are able to observe following the outbreak. Although the production of weaned pigs recovered to a level that is not significantly different (P < 0.01) from the baseline, we cannot definitively declare that there was no effect beyond week t + 35. Nonetheless, it appears that any continued effect is likely to be very small relative to the large effect occurring before week t + 35.

We detected a consistent decrease in production until the 5th week after the outbreak report, followed by a non-monotonic recovery. All performance parameters followed a similar non-monotonic recovery pattern. Each indicator manifested a sharp worsening after the outbreak, followed by partial recovery and at least one mild period of deterioration. The dynamic up-and-down impact of PRRS on weaned pig production was surprising. The precise causes are unclear, but the disease may spread more slowly and unevenly through the sow herd than anticipated, particularly on large units with multiple cohorts, in addition to possible incoming flows of replacement sows. This effect might also explain the longer period of recovery in our study, versus another study that found production returned to the baseline in 16.5 weeks for cohorts vaccinated with an MLV and using herd closure as a control strategy (21).

Other performance indicators provided consistent signals. Pre-weaning mortality increased sharply in weeks t − 1 to t + 1, declined to pre-outbreak levels by t + 10, and then oscillated about that level until about t + 24. Sow mortality increased in week t + 1 and remained above baseline levels until week t + 5. The increase in sow mortality could affect the age structure of the herd and consequently its production. Stillbirths increased until week t + 12, indicating that some infected sows carried damaged fetuses to birth. The number of stillbirths remained elevated through t + 36, suggesting that infected sows may have a higher probability of producing stillborn piglets for more than one pregnancy. The failure to conceive was followed by repeated services, which must have contributed to the lag in weaned pig production in later weeks. The numbers of pigs aborting or dying indicated that PRRS had its strongest effects on fetuses. PRRS kills relatively few sows and piglets, though the economic damage from sow mortality and/or their subsequently reduced productivity is important.

Information regarding the strains of PRRS virus that affected each farm was not available for this study, as systematic sequencing of PRRS virus following outbreaks is still scarce. More than one strain might affect a given area, although in general genetic variation is more related to temporal rather than spatial variation (10). Using a sample of 16 farms may help capture the variability of PRRS outbreaks in the industry, assuming different strains may be affecting different farms.

According to a number of studies, no vaccine prevents PRRS infection, but vaccination may reduce the risk of infection and may also reduce the intensity of outbreaks by reducing the amount of virus excreted by ill animals (2631). Therefore, our results may show smaller damages than those that would be obtained for farms that do not vaccinate. Similarly, because the farms analyzed in this study belong to a firm with standardized protocols for disease management, our measure of PRRS’ impact could be smaller than would be measured on farms with poorer protocols.

We developed and used a straightforward approach to quantify the dynamic effect of PRRS on weaned pig production within sow farms. We found that PRRS decreased weaned pig production for at least 35 weeks among the firms studied. The magnitude of PRRS’ impact, as expressed in the duration and magnitude of the output decline, were both greater than anticipated. We found that recovery oscillated about a rising trend, i.e., recovery does not depict a clear monotonic increase in production, suggesting that farms suffered from a continuing circulation of the disease within the herd and/or a lingering effect on sows and piglets. Analysis of the underlying performance indicators provided additional insight regarding how PRRS affects farm output over time. Previous studies have utilized numerous assumptions to develop estimates of the total annualized losses to the swine industry due to PRRS (7, 9). We have not attempted to replicate those studies. However, our results suggest PRRS may cause significantly higher losses on sow farms than has been estimated previously. Further, we believe that the losses identified in our farm sample are likely to be smaller than those on the average sow farm infected with PRRS. Nonetheless, we found substantial variation in performance among even a set of relatively standardized 16 farms. There is thus need for caution when using simple averages, as we often have done, rather than distributions across farms.

## Author Contributions

All the authors have met the four criteria described in the guidelines: PV-D designed analyses and data interpretation, revised and approved the version to be published, and agreed to be accountable for all aspects of the work. LJ, JA, RM, and AP revised and approved the version to be published, and agreed to be accountable for all aspects of the work.

## Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

## Acknowledgments

The Becas-Chile program of the National Commission for Scientific and Technological Research (CONICYT) has supported PV-D to complete his MS and PhD degrees. Additional support for this research was provided by the University of Minnesota MnDrive. LS acknowledges support from the National Institute of Food and Agriculture (NIFA). We thank the swine firm for sharing their data from the 16 sow farms.

## Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fvets.2018.00102/full#supplementary-material

## References

1. Dougherty C. Introduction to econometrics. Oxford, NY: Oxford University Press (2011).

2. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis. Boston, MA: John Wiley & Sons (2011).

3. Hennessy DA. Behavioral Incentives, Equilibrium Endemic Disease, and Health Management Policy for Farmed Animals. Am J Agric Econ (2007) 89(3):698–711. doi: 10.1111/j.1467-8276.2007.01001.x

4. Jarvis LS, Valdes-Donoso P. A Selective Review of the Economic Analysis of Animal Health Management. J Agric Econ (2018) 69(1):201–25. doi: 10.1111/1477-9552.12131

5. Albina E. Epidemiology of porcine reproductive and respiratory syndrome (PRRS): an overview. Vet Microbiol (1997) 55(1-4):309–16. doi: 10.1016/S0378-1135(96)01322-3

6. Holck JT, Polson DD. The financial impact of PRRS virus. PRRS Compendium Producer Edition. National Pork Board (2003). p. 47–54.

7. Neumann EJ, Kliebenstein JB, Johnson CD, Mabry JW, Bush EJ, Seitzinger AH, et al. Assessment of the economic impact of porcine reproductive and respiratory syndrome on swine production in the United States. J Am Vet Med Assoc (2005) 227(3):385–92. doi: 10.2460/javma.2005.227.385

8. Lunney JK, Benfield DA, Rowland RR. Porcine reproductive and respiratory syndrome virus: an update on an emerging and re-emerging viral disease of swine. Virus Res (2010) 154(1-2):1–6. doi: 10.1016/j.virusres.2010.10.009

9. Holtkamp DJ, Kliebenstein JB, Neumann EJ, Zimmerman JJ, Rotto HF, Yoder TK. Assessment of the economic impact of porcine reproductive and respiratory syndrome virus on United States pork producers. J Swine Health Prod (2013) 21:72–84.

10. Perez AM, Davies PR, Goodell CK, Holtkamp DJ, Mondaca-Fernández E, Poljak Z, et al. Lessons learned and knowledge gaps about the epidemiology and control of porcine reproductive and respiratory syndrome virus in North America. J Am Vet Med Assoc (2015) 246(12):1304–17. doi: 10.2460/javma.246.12.1304

11. Pitkin A, Deen J, Dee S. Further assessment of fomites and personnel as vehicles for the mechanical transport and transmission of porcine reproductive and respiratory syndrome virus. Can J Vet Res (2009) 73(4):298–302.

12. Martín-Valls GE, Kvisgaard LK, Tello M, Darwich L, Cortey M, Burgara-Estrella AJ, et al. Analysis of ORF5 and full-length genome sequences of porcine reproductive and respiratory syndrome virus isolates of genotypes 1 and 2 retrieved worldwide provides evidence that recombination is a common phenomenon and may produce mosaic isolates. J Virol (2014) 88(6):3170–81. doi: 10.1128/JVI.02858-13

13. Valdes-Donoso P, Jarvis LS, Wright D, Alvarez J, Perez AM. Measuring Progress on the Control of Porcine Reproductive and Respiratory Syndrome (PRRS) at a Regional Level: The Minnesota N212 Regional Control Project (Rcp) as a Working Example. PLoS One (2016) 11(2):e0149498. doi: 10.1371/journal.pone.0149498

14. Dee S, Otake S, Oliveira S, Deen J. Evidence of long distance airborne transport of porcine reproductive and respiratory syndrome virus and Mycoplasma hyopneumoniae. Vet Res (2009) 40(4):39. doi: 10.1051/vetres/2009022

15. Otake S, Dee S, Corzo C, Oliveira S, Deen J. Long-distance airborne transport of infectious PRRSV and Mycoplasma hyopneumoniae from a swine population infected with multiple viral variants. Vet Microbiol (2010) 145(3-4):198–208. doi: 10.1016/j.vetmic.2010.03.028

16. Silva GS, Linhares DC. Assessing anti-PRRSV biosecurity. 48th Annual Meeting of the American Association of Swine Veterinarians. (2017). p. 27–30.

17. Zimmerman JJ, Yoon KJ, Wills RW, Swenson SL. General overview of PRRSV: a perspective from the United States. Vet Microbiol (1997) 55(1-4):187–96. doi: 10.1016/S0378-1135(96)01330-2

18. Dewey CE, Johnston WT, Gould L, Whiting TL. Postweaning mortality in Manitoba swine. Can J Vet Res (2006) 70(3):161–7.

19. Alvarez J, Sarradell J, Kerkaert B, Bandyopadhyay D, Torremorell M, Morrison R, et al. Association of the presence of influenza A virus and porcine reproductive and respiratory syndrome virus in sow farms with post-weaning mortality. Prev Vet Med (2015) 121(3-4):240–5. doi: 10.1016/j.prevetmed.2015.07.003

20. Torremorell M, Henry S, Christianson WT. “Eradication using herd closure,”. In: Zimmerman J, Yoon KJ, Neumann E, editors. PRRS compendium. National Pork Board (2003). p. 111–5.

21. Linhares DC, Cano JP, Torremorell M, Morrison RB. Comparison of time to PRRSv-stability and production losses between two exposure programs to control PRRSv in sow herds. Prev Vet Med (2014) 116(1-2):111–9. doi: 10.1016/j.prevetmed.2014.05.010

22. Holtkamp DJ, Polson DD, Torremorell M, Morrison B, Classen DM, Becton L, et alAmerican Association of Swine Veterinarians, United States Department of Agriculture PRRS-Coordinated Agricultural Project, et al. [Terminology for classifying the porcine reproductive and respiratory syndrome virus (PRRSV) status of swine herds]. Tierarztl Prax Ausg G Grosstiere Nutztiere (2011) 39(2):101–12.

23. Angrist JD, Pischke J-S. Mostly harmless econometrics: An empiricist’s companion. UK: Princeton University Press (2008).

24. Statacorp. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP (2013).

25. Schulz L. Ag Decision Maker: Historic Hog and Lamb Prices. Ames, Iowa: Iowa State University (2016).

26. Benson JE, Yaeger MJ, Lager KM. Effect of porcine reproductive and respiratory syndrome virus (PRRSV) exposure dose on fetal infection in vaccinated and nonvaccinated swine. Swine Health and Production (2000) 8:155–60.

27. Scortti M, Prieto C, Simarro I, Castro JM. Reproductive performance of gilts following vaccination and subsequent heterologous challenge with European strains of porcine reproductive and respiratory syndrome virus. Theriogenology (2006) 66(8):1884–93. doi: 10.1016/j.theriogenology.2006.04.043

28. Cano JP, Dee SA, Murtaugh MP, Pijoan C. Impact of a modified-live porcine reproductive and respiratory syndrome virus vaccine intervention on a population of pigs infected with a heterologous isolate. Vaccine (2007) 25(22):4382–91. doi: 10.1016/j.vaccine.2007.03.031

29. Zuckermann FA, Garcia EA, Luque ID, Christopher-Hennings J, Doster A, Brito M, et al. Assessment of the efficacy of commercial porcine reproductive and respiratory syndrome virus (PRRSV) vaccines based on measurement of serologic response, frequency of gamma-IFN-producing cells and virological parameters of protection upon challenge. Vet Microbiol (2007) 123(1-3):69–85. doi: 10.1016/j.vetmic.2007.02.009

30. Geldhof MF, Vanhee M, van Breedam W, van Doorsselaere J, Karniychuk UU, Nauwynck HJ. Comparison of the efficacy of autogenous inactivated Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) vaccines with that of commercial vaccines against homologous and heterologous challenges. BMC Vet Res (2012) 8:182. doi: 10.1186/1746-6148-8-182

31. Zhang H, Kono H, Kubota S. An integrated epidemiological and economic analysis of vaccination against highly pathogenic Porcine Reproductive and Respiratory Syndrome (PRRS) in Thua Thien Hue Province, Vietnam. Asian-australas J Anim Sci (2014) 27(10):1499–512. doi: 10.5713/ajas.2014.14060

32. Polson D, Marsh W, Dial G, Christianson W. Financial impact of porcine epidemic abortion and respiratory syndrome (PEARS). Proceedings of the International Pig Veterinary Society Congress. (1992). p. 47–54.

Keywords: Porcine Reproductive and Respiratory Syndrome, Production Impacts, Sow Farms, US Swine Industry, Endemic Animal Disease, Fixed effects model

Citation: Valdes-Donoso P, Alvarez J, Jarvis LS, Morrison RB and Perez AM (2018). Production Losses From an Endemic Animal Disease: Porcine Reproductive and Respiratory Syndrome (PRRS) in Selected Midwest US Sow Farms. Front. Vet. Sci. 5:102. doi: 10.3389/fvets.2018.00102

Received: 27 October 2017; Accepted: 23 April 2018;
Published: 16 May 2018

Edited by:

Salome Dürr, Universität Bern, Switzerland

Reviewed by:

Krishna Thakur, University of Prince Edward Island, Canada
Jarkko Niemi, Natural Resources Institute Finland (Luke), Finland

Copyright © 2018 Valdes-Donoso, Alvarez, Jarvis, Morrison and Perez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Pablo Valdes-Donoso, pvaldesdonoso@ucdavis.edu

Present Address: Agricultural Issues Center, University of California, Davis, CA, United States