Prevalence and risk factors for infection with Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) in swine herds in Illinois (USA)

Infection of swine herds with Porcine Reproductive and Respiratory Syndrome virus (PRRSV) can cause significant economic losses. There have been few epidemiologic studies of PRRSV. Little is known about the risk factors for infection of swine herds with PRRSV. The objectives of this study were to use an existing (1992) serum bank of samples from Illinois (USA) swine herds to estimate the prevalence of PRRSV infection, and to identify risk factors for PRRSV infection and disease in Illinois swine farms. It was hypothesized that the risk of PRRSV infection and disease would increase with (1) increasing herd size, (2) total confinement housing, (3) a greater number of purchases of gilts and boars, (4) purchase of semen for artificial insemination, and (5) not isolating boars and gilts after purchase. In 1992, serum samples from sows on 176 swine farms were obtained from the Illinois state diagnostic laboratories as part of the state pseudorabies testing program. Thereafter these samples were kept frozen at $-20 \,^\circ$C in a serum data bank. During 1997-98 these samples were thawed and tested for antibodies to the PRRS virus, using the USDA-approved IDEXX PRRS ELISA test. Telephone interviews on demographic, management, and housing characteristics had been obtained for 140 of these farms in 1992. Data relevant to testing of hypotheses about the risk of PRRSV infection included the following: (1) number of boars purchased in the last year, (2) number of gilts purchased in the last year, (3) average number of sows in the herd in the last year, and (4) type of housing (total confinement, partial confinement, outdoor lots, pasture). During 1997, 103 of these farms (74%) were re-contacted by telephone and an updated survey was conducted, specifically targeted towards obtaining information on PRRS and its risk factors. Of the 103 farms for which herd survey information was available for data analysis, 50 (49%) had been diagnosed with PRRSV infection between 1992 and 1997. In 37 (74%) of the herds diagnosed with PRRSV infection, there were clinical signs, mostly epidemic abortions (62%), pre-weaning mortality (36%) and respiratory signs in nursery (58%) and growing-finishing (52%) pigs. There were PRRSV serological test results on at least 20 sows for 54 of the surveyed herds. There was a mean of 34 serum samples tested per herd (median = 30; range: 20-76), with a mean seroprevalence of 28% (median = 10%), with 36 of the herds (67%) having at least one seropositive sow. Bivariate analysis of data from the herd surveys indicated that increased risk of PRRS diagnosis was associated with not isolating gilts after purchase (OR = 6.2; $\rm p < 0.001$), a higher number of sows (mean = 351 vs. 181 with no PRRS diagnosis; $\rm p < 0.001$), total confinement (OR = 3.0; $\rm p < 0.01$), purchasing semen for artificial insemination (OR = 2.5; p = 0.02), and more gilts purchased (mean = 40 vs. 17 with no PRRS diagnosis; p = 0.03). Similar results were obtained when the outcome analyzed was PRRS disease. Multiple logistic regression analysis indicated that an increased risk of PRRS diagnosis was associated with a higher number of sows (adjusted OR = 1.003; $\rm p < 0.01$), not isolating gilts after purchase (aOR = 3.3; p = 0.02), and purchase of semen for artificial insemination (aOR = 2.2; p = 0.04). Multiple linear regression analysis of risk factors for sow PRRSV seroprevalence in herds with survey results indicated only that the purchase of boars increased this risk $\rm (p < 0.01)$. Our conclusions are the following