Spatial distribution of the exposure to C. burnetii in the Netherlands: measuring and modelling airborne transmission and combining transmission data with a dose response model using local data on the incidence of Q-fever

Between 2007 and 2010, large community outbreaks of Q fever have occurred in the Netherlands, with over 3500 recognized cases. Two years before the first documented human outbreak in 2007, abortion waves were reported on a large goat farm in the most affected region. This pointed towards dairy goat farms as potential sources for human Q fever infection, supported later by an epidemiological outbreak analysis. The contribution of various transmission routes of Q fever is not well understood, but infection by inhalation of contaminated aerosols is thought to be the primary transmission route. At present, control of Q fever is based on (partly) untested hypotheses on sources and mechanisms of transmission. It is assumed that after partition sheep placenta and body fluids contain bacteria and end up in deep litter systems. Manure is spread on the field and dries, hibernating bacteria scatter via dust particles, thus resulting in exposure of animals and humans. However, little is known about sources of C. burnetii on the farm level and transmission and contributing factors such as farm management, farm design, hygiene measures. The short term effect of the extremely rigorous culling strategy, involving all goats with lambs on farms with a milk tank testing positive for Q fever, is visible but likely to be incomplete. The number of human cases has been reduced in 2010 as compared to previous years, but new cases are still appearing. Evidence based decision rules for public health interventions in case of future outbreaks of Q fever are urgently needed. What interventions are appropriate to limit the risk of Q fever for the general population? How can we apply measures at goat farms in order to forestall transmission to the environment, other farms, or inhabited areas? Is large scale culling (as applied in the Netherlands in 2009) a sensible approach or do alternatives or combinations of alternatives exist? How long do exposed environments continue to cause increased risks? Should vaccination be applied, or are hygiene measures or improvements in ventilation techniques more promising? At present, control of Q fever is based on (partly) untested hypotheses on sources and mechanisms of transmission. The project has the following aims: 1. measure exposure to C. burnetii with advanced aerosol capturing techniques, in combination with molecular detection methods in air and relating air levels to different activities on the farm, farm characteristics and management practices using advanced technologies for sampling and analysis; 2. model exposure to C. burnetii in the environment including emission, dispersion and local exposure of Q fever. Modeling will focus on farm to human transmission and results will be used to describe exposure risk charts; 3. model exposure to C. burnetii using transmission kernel calculations, which can also be translated in risk charts; 4. perform a quantitative exposure or dose response analysis for antibody response after infection and notification data to reconstruct exposure risk patterns; 5. combine the quantitative dose-response analysis with knowledge about the kinetics of the antibody response after infection, and the geographical distribution of serological profiles and notification data to define decision rules for risk groups such as pregnant women and excluding donors from donating blood. 6. Combine exposure and dose-response models to estimate geographical distribution of risks and compare with observed serological and clinical data. The project aims at integrating different approaches, and works from different angles at the above questions. By using C. burnetii positive farm geo-distribution data with emission estimates as input for physical dispersion modeling and transmission kernel calculations, to estimate transmission and exposure of humans. The generated exposure data in the form of exposure risk maps with be related to case clustering and validated by measurement of C. burnetii in air and the environment. The relationship will be made by comparing estimated Q fever occurrence on the basis of described exposure response models available for laboratory animals and humans, with observed Q fever occurrence. At the same time, the exposure distribution will be estimated through exposure or dose response information from epidemiological survey data, starting from time-space clustering of Q fever occurrence. So, risk maps will be produced in different ways. By working from these different ends, integrated approaches will be developed and results will be cross-validated. Results will be disseminated by development of a generally available measurement protocol for C. burnetii, and a series of scientific papers. Results will be communicated with experts in the field and policy makers during a worshop. An attempt will be made to translate results of this workshop into a consensus statement.