Experimental intrauterine infection with Prevotella bivia in New Zealand White rabbits

Objective: The purpose of this study was to develop a model of chronic intrauterine and fetal infection with Prevotella bivia, an anaerobe of the lower genital tract that is associated often with bacterial vaginosis. Study design: Thirty timed pregnant New Zealand White rabbits on gestational day 21 were inoculated with P bivia or saline solution in a planned ratio of 4:1 (24 P bivia:6 saline solution). Rabbits were inoculated 6 cm transcervically with 10⁵ to 10⁸ colony-forming units/uterine horn of P bivia or with saline solution. Necropsy was scheduled on days 4,6, or 7 after inoculation. Cultures were collected from blood, uterus, amniotic fluid and fetal brain, lung, and heart. Tissues from placenta, uterus, fetal brain, and lung were evaluated with the histologic inflammation score, with a range of 0 to 13. Amniotic fluid was assayed for tumor necrosis factor-α by bioassay. Animals with contamination by other organisms were excluded. Categoric data were evaluated with the use of the Fisher exact test, and continuous data were evaluated with the use of the Wilcoxon rank sum. Results: After the exclusion of 8 animals because of contamination with other organisms, 22 animals were evaluated. Of 3 rabbits with an inoculum of 10⁸ P bivia colony-forming units/horn, 2 animals (67%) had fever within 24 hours. These results were not compatible with chronic, subclinical infection. Therefore, 14 does had inocula of 10^5-6 P bivia colony-forming units/horn, with necropsy planned at day 4 (n = 5 animals), day 6 (n = 3 animals), and day 7 (n = 6 animals), and 5 animals were inoculated with saline solution. Animals that had been inoculated with P bivia were significantly more likely to have a positive culture than were those animals that were inoculated with saline solution (64% vs 0%; P<.04). Preterm delivery without fever occurred in 21% of does (3/14 does) that were inoculated with P bivia overall and in 33% of the does (3/9 does) that were followed for 6 to 7 days. No saline-solution inoculated animal had preterm birth. There was an increase in amniotic fluid tumor necrosis factor-α levels over time in the P bivia group (P = .12). Histologic inflammation scores were not significantly different between P bivia and saline solution groups. Conclusion: Inoculation with P bivia at 10 ^5-6 colony-forming units/horn leads to chronic intrauterine and fetal infection that are accompanied by preterm birth in up to 33% of cases. This model may serve to explore the mechanism of preterm birth that is induced by chronic infection with genital tract anaerobes.