A variety of methods for eliminating the PRRS virus from pig production sites have been successfully applied. However, success in maintaining a PRRS virus-free status for extended periods of time following elimination has been inconsistent and unpredictable. The objective of this study was to evaluate whether risks measured using version 1 of the American Association of Swine Veterinarians (AASV) PRRS Risk Assessment for the Breeding Herd, season of year and method by which swine breeding herd sites were established PRRS virus-free were associated with how long they retained their virus-free status. Thirty-three swine farrow-to-wean breeding herd sites that were established as PRRS virus-free by either populating a new site with virus-free breeding animals or by completely depopulating the site and repopulating with PRRS virus-free breeding animals were enrolled in this study. Survival analysis, using the Cox proportional hazards model and Kaplan-Meier survival curves, was performed where the outcome was the duration of time PRRS virus-free breeding herd sites remained virus-free ("survived"). Covariates evaluated included the internal and external risk scores measured by the PRRS Risk Assessment for the Breeding Herd as well as the season and the method by which the site was established free of the PRRS virus. All but 5 (15%) of the 33 sites became positive to the PRRS virus during the course of the study and approximately 40% became positive within 1 year from when they were established free of the PRRS virus. A higher external risk score was associated with a greater risk of becoming positive to the PRRS virus and shorter survival times. The internal risk score was not significantly associated with survival. Establishing breeding herd sites free of the PRRS virus in winter months (November through February) was associated with a greater risk of becoming positive to the PRRS virus and shorter survival times compared to those established in non-winter months. The association between the risk of becoming positive to the PRRS virus and the external risk score was confounded by the method the site was established PRRS virus-free.