Transplacental Transmission of Bluetongue Virus 8 in Cattle, UK

To determine whether transplacental transmission could explain overwintering of bluetongue virus in the United Kingdom, we studied calves born to dams naturally infected during pregnancy in 2007–08. Approximately 33% were infected transplacentally; some had compromised health. In all infected calves, viral load decreased after birth; no evidence of persistent infection was found.

negative PCRs for bovine viral diarrhea virus (S.W., pers. comm.). Although calf X died of colisepticemia, this illness probably resulted from the calf's weakness and inability to consume colostrum. No infectious cause for the early postnatal death of calf Y, other than bluetongue, was identifi ed; pathologic fi ndings for calves 13 and 29 are described elsewhere (S.W. et al., unpub. data). Calf 27, which had negative BTV test results, was born with hypermobility of the fetlock joints, unilateral carpal valgus, and arthrogryposis. All other calves were reported to be healthy.
Time windows for possible in utero infection of each calf were calculated according to the BTV testing history of the dam and the birth date of the calf (Figure). These windows were used to investigate effect of stage of gestation on the probability of transplacental transmission. To account for uncertainty in the date of infection, we used Bayesian methods (online Technical Appendix, available from www.cdc.gov/EID/content/15/12/2025-Techapp.pdf). The probability of transplacental transmission increased with the time of gestation during which the dam became infected (β 1 0.033; 95% credibility interval 0.014-0.063).

Conclusions
This detailed fi eld study, which combines data on BTV infection in cows with data on transplacentally acquired infection in their offspring, demonstrates that the BTV-8 strain circulating in northern Europe can cross the bovine placenta in a high proportion (33%) of cases and infect calves when dams are infected during pregnancy. A similar study in continental Europe suggested a rate of ≈10% (13). However, because the transmission season was longer in some of these countries, many seropositive dams could have been infected before pregnancy, leading to underestimation of the probability of transplacental infection. In our study, we tested only calves from dams infected between August and December 2007 and known to be pregnant at the time of infection. Furthermore, analysis of our data suggests that transplacental transmission is more likely when infection occurs later in gestation; indeed, most of the dams in this study would have been in the second or third gestation trimester when infected ( Figure), which may have increased our estimated rate over that found in continental Europe.
Transplacental transmission is of particular concern for policy makers because it may result in the birth of immune-tolerant, persistent carriers, as has happened with bovine viral diarrhea virus (14). In our study, all BTV-positive calves other than X and Y were tested after they had received colostrum and, hence, maternal antibodies. The presence of BTV antibodies in calf Y suggests that fetal antibody formed in response to in utero infection, yet calf X had no detectable antibodies against BTV despite strongly positive rRT-PCR results. Calf X was infected late in gestation ( Figure), when it should have been capable of mounting its own antibody response (15). Antibody-negative PCR-positive calves have been reported elsewhere (13). Follow-up testing is needed to assess whether such calves remain persistently infected; however, because calf X died a few days after birth, follow-up testing was not possible.
RNA declined in all retested calves (Table); most were PCR-negative by the end of the study, including dummy calf 13. Therefore, our results do not suggest that transpla-  cental infection with BTV-8 results in subclinical, persistent carriers. Nonetheless, the fi nding that some calves may be born with deformaties after the virus has cleared may lead to underestimation of the economic effects of BTV; calf 27, which was born with limb deformities to a BTV positive dam, could be such a case. Live virus has been successfully isolated from only 4 transpacentally infected calves (including calf X described in this study), all of which received either no maternal colostrum or only pooled colostrum (9,13). Further work is needed to assess whether infectious virus can be isolated from healthy transplacentally infected calves that have colostrum-derived maternal antibodies, because infectious virus needs to be present if transplacental infection is to play a major role in overwintering. In conclusion, future emerging BTV strains should be considered to have the potential for transplacental transmission until investigations show otherwise.

Acknowledgments
We are indebted to all the farmers who participated in this study for their invaluable cooperation. We also thank many colleagues at the Institute for Animal Health, Pirbright, the Animal Health divisional offi ces at Bury St. Edmunds and Chelmsford, and the regional laboratories of the Veterinary Laboratories Agency (VLA) at Bury St. Edmunds and Winchester for all their help and guidance. As well, we thank Simon Carpenter, Christopher Sanders, James Barber, Anthony Greenleaves, and Alan Hurst for their support and contributions to this study.
This fi eld study, led by the Institute for Animal Health, Pirbright, in cooperation with Animal Health through their divisional offi ces at Bury St. Edmund and Chelmsford, and the VLA through where B i is the date of birth of calf i, ( ) L i I is the earliest date of infection (either 2 days before the last negative PCR result, if available (see Table 1 and Figure 1

Transplacental Infection in Relation to Time of Gestation and Pregnancy
The probability of transplacental transmission from dam to calf for the ith calf-dam pair is given by where t i is the stage of gestation at which the ith dam was infected. A Bayesian approach assuming non-informative (diffuse Normal) priors was used to estimate the parameters (the where are the earliest and latest times during gestation at which the dam could have been affected, respectively (defined in equation (1)). The final model for the probability of transplacental transmission was constructed starting from a linear function (m = 1) in equation (1) and sequentially adding higher-order terms (m = 2,3,...) until there was no improvement in model fit, as judged by the Deviance Information Criterion (DIC) (4).
Multiple chains were run to check convergence and, in each case, estimates were based on 50,000 iterations of the chain, with the preceding 10,000 iterations discarded.