Epidemiology of Q fever in dairy goat herds in the Netherlands

Between 2007 and 2009, the largest human Q fever epidemic ever described occurred in the Netherlands. The source was traced back to dairy goat farms, where abortion storms caused by Coxiella burnetii had been observed. Intervention measures included vaccination of dairy goats, followed by one-time culling of all pregnant animals from infected dairy goat herds in early 2010. These measures aimed to prevent further shedding of C. burnetii into the environment during the next kidding season, thus minimizing human exposure. Beginning in 2010, the number of human cases sharply decreased, suggesting that the intervention measures, including continued yearly vaccination of all dairy goat herds, were effective. One putative cause of the abortion storms in dairy goat herds was the intensive husbandry systems in which the goats were, and still are, kept. However, as this thesis shows, the abortion storms in Dutch dairy goat herds could not be fully simulated by demographic factors, such as herd size or reproductive patterns. Aspects of the pathogenesis in goats probably play and unknown but essential role in the observed abortion patterns. There are gaps in our current understanding of the drivers of Q fever abortion storms, and this may hamper the prevention of future problems on intensive goat farms and in humans living in their environment in a non-vaccination situation. Transmission via inhalation is considered the main mode of transmission for C. burnetii, but at the time of the Dutch epidemic, only a few published studies had actually demonstrated airborne C. burnetii. A method to detect C. burnetii DNA in airborne dust samples of size fractions that can be inhaled by humans is described in this thesis. With this method, C. burnetii DNA was indeed detected in airborne dust samples collected at three affected dairy goat farms. This finding supports the general assertion that airborne transmission is a likely route of exposure. As this thesis shows, vaccination reduces bacterial shedding by small ruminants. Both prevalence and bacterial load was reduced in the uterine fluid, vaginal mucus and milk of animals from vaccinated herds compared with animals from unvaccinated herds. These effects were most pronounced in animals during their first pregnancy. Results indicate that vaccination may reduce bacterial load in the environment and thus lessen human exposure to C. burnetii. Animals with a positive PCR in uterine fluid can be considered high risk animals, because they may shed massive amounts of bacteria during delivery. In a field setting, sampling of uterine fluid is not possible. To predict which animals are high risk animals, test characteristics of current tests for C. burnetii were evaluated for vaginal mucus (PCR), milk (PCR and ELISA), in recently infected unvaccinated herds. Due to poor test performance, identification and selective removal of high risk animals only (the test and cull strategy) to protect public health seems currently unfeasible. Because emergency vaccination of these animals is also not effective, the hazard of massive bacterial shedding during abortion or parturition can only be prevented by removal of all pregnant animals from recently infected naive (i.e. non vaccinated) herds.