Human Q fever incidence is associated to spatiotemporal environmental conditions

J P G Van Leuken; A N Swart; J Brandsma; W Terink; J Van de Kassteele; P Droogers; F Sauter; A H Havelaar; W Van der Hoek

doi:10.1016/j.onehlt.2016.03.004

Human Q fever incidence is associated to spatiotemporal environmental conditions

One Health. 2016 Apr 4:2:77-87. doi: 10.1016/j.onehlt.2016.03.004. eCollection 2016 Dec.

Authors

J P G Van Leuken^{1

2}, A N Swart¹, J Brandsma³, W Terink³, J Van de Kassteele¹, P Droogers³, F Sauter⁴, A H Havelaar^{1

2

5}, W Van der Hoek¹

Affiliations

¹ Centre for Infectious Disease Control (CIb), National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands.
² Institute for Risk Assessment Sciences (IRAS), Faculty of Veterinary Medicine, Utrecht University, Utrecht, The Netherlands.
³ Future Water, Wageningen, The Netherlands.
⁴ Environmental Safety (M&V), National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands.
⁵ Emerging Pathogens Institute, University of Floriday, Gainesville, Florida, United States.

Abstract

Airborne pathogenic transmission from sources to humans is characterised by atmospheric dispersion and influence of environmental conditions on deposition and reaerosolisation. We applied a One Health approach using human, veterinary and environmental data regarding the 2009 epidemic in The Netherlands, and investigated whether observed human Q fever incidence rates were correlated to environmental risk factors. We identified 158 putative sources (dairy goat and sheep farms) and included 2339 human cases. We performed a high-resolution (1 × 1 km) zero-inflated regression analysis to predict incidence rates by Coxiella burnetii concentration (using an atmospheric dispersion model and meteorological data), and environmental factors - including vegetation density, soil moisture, soil erosion sensitivity, and land use data - at a yearly and monthly time-resolution. With respect to the annual data, airborne concentration was the most important predictor variable (positively correlated to incidence rate), followed by vegetation density (negatively). The other variables were also important, but to a less extent. High erosion sensitive soils and the land-use fractions "city" and "forest" were positively correlated. Soil moisture and land-use "open nature" were negatively associated. The geographical prediction map identified the largest Q fever outbreak areas. The hazard map identified highest hazards in a livestock dense area. We conclude that environmental conditions are correlated to human Q fever incidence rate. Similar research with data from other outbreaks would be needed to more firmly establish our findings. This could lead to better estimations of the public health risk of a C. burnetii outbreak, and to more detailed and accurate hazard maps that could be used for spatial planning of livestock operations.

Keywords: Atmospheric dispersion model; Concentration; Coxiella burnetii; Environmental; Multivariate; Prediction; Q fever; Risk factors; Soil moisture; Vegetation.