Research and Innovation Opportunities to Improve Epidemiological Knowledge and Control of Environmentally Driven Zoonoses.

While zoonotic diseases are defined by transmission processes between animals and humans, for many of these diseases the presence of a contaminated environmental source is the cause of transmission. Most zoonoses depend on complex environmentally driven interactions between humans and animals, which occur along an occupational and recreational environmental continuum, including farming and animal marketing systems, environmental management systems, and community leisure environments. Environmentally driven zoonoses (EDZs) are particularly challenging to diagnose and control as their reservoirs are in the natural environment and thus often escape conventional surveillance systems that rely on host monitoring. Changes in the environment as a result of climate change [1], human population density [2], and intensification of agriculture [3] have been linked to increasing transmission events for this group of infections. As such, there is a recognised need to be able to detect the presence of EDZs in the environment as a means to better anticipate transmission events and improve source attribution investigations. Finally, the recognition that a One Health approach is needed to combat these infections is signalling to governments the need to develop policy that optimises trade-offs across human, animal, and environmental health sectors. In this review, we discuss and critically appraise the main challenges relating to the epidemiology, diagnosis, and control of environmental zoonotic disease. Using a set of exemplar diseases, including avian influenza and antimicrobial resistant pathogens, we explore the epidemiological contexts (risk factors) within which these infections not only impact human health but also contribute to animal health and environmental impacts. We then critically appraise the surveillance challenges of monitoring these infections in the environment and examine the policy trade-offs for a more integrated approach to mitigating their impacts.

Profiling Risk Factors for Household and Community Spatiotemporal Clusters of Q Fever Notifications in Queensland between 2002 and 2017.

Q fever, caused by the bacterium Coxiella burnetii, is an important zoonotic disease worldwide. Australia has one of the highest reported incidences and seroprevalence of Q fever, and communities in the state of Queensland are at highest risk of exposure. Despite Australia's Q fever vaccination programs, the number of reported Q fever cases has remained stable for the last few years. The extent to which Q fever notifications cluster in circumscribed communities is not well understood. This study aimed to retrospectively explore and identify the spatiotemporal variation in Q fever household and community clusters in Queensland reported during 2002 to 2017, and quantify potential within cluster drivers. We used Q fever notification data held in the Queensland Notifiable Conditions System to explore the geographical clustering patterns of Q fever incidence, and identified and estimated community Q fever spatiotemporal clusters using SatScan, Boston, MA, USA. The association between Q fever household and community clusters, and demographic and socioeconomic characteristics was explored using the chi-squared statistical test and logistic regression analysis. From the total 2175 Q fever notifications included in our analysis, we found 356 Q fever hotspots at a mesh-block level. We identified that 8.2% of Q fever notifications belonged to a spatiotemporal cluster. Within the spatiotemporal Q fever clusters, we found 44 (61%) representing household clusters and 20 (27.8%) were statistically significant with an average cluster size of 3 km radius. Our multivariable model shows statistical differences between cases belonging to clusters in comparison with cases outside clusters based on the type of reported exposure. In conclusion, our results demonstrate that clusters of Q fever notifications are temporally stable and geographically circumscribed, indicating a persistent common exposure. Furthermore, within individuals in household and community clusters, abattoir exposure (a traditional occupational exposure) was rarely reported by individuals.

Near-term forecasting of companion animal tick paralysis incidence: An iterative ensemble model.

Tick paralysis resulting from bites from Ixodes holocyclus and I. cornuatus is one of the leading causes of emergency veterinary admissions for companion animals in Australia, often resulting in death if left untreated. Availability of timely information on periods of increased risk can help modulate behaviors that reduce exposures to ticks and improve awareness of owners for the need of lifesaving preventative ectoparasite treatment. Improved awareness of clinicians and pet owners about temporal changes in tick paralysis risk can be assisted by ecological forecasting frameworks that integrate environmental information into statistical time series models. Using an 11-year time series of tick paralysis cases from veterinary clinics in one of Australia's hotspots for the paralysis tick Ixodes holocyclus, we asked whether an ensemble model could accurately forecast clinical caseloads over near-term horizons. We fit a series of statistical time series (ARIMA, GARCH) and generative models (Prophet, Generalised Additive Model) using environmental variables as predictors, and then combined forecasts into a weighted ensemble to minimise prediction interval error. Our results indicate that variables related to temperature anomalies, levels of vegetation moisture and the Southern Oscillation Index can be useful for predicting tick paralysis admissions. Our model forecasted tick paralysis cases with exceptional accuracy while preserving epidemiological interpretability, outperforming a field-leading benchmark Exponential Smoothing model by reducing both point and prediction interval errors. Using online particle filtering to assimilate new observations and adjust forecast distributions when new data became available, our model adapted to changing temporal conditions and provided further reduced forecast errors. We expect our model pipeline to act as a platform for developing early warning systems that can notify clinicians and pet owners about heightened risks of environmentally driven veterinary conditions.

Geographical Variation in Coxiella burnetii Seroprevalence in Dairy Farms Located in South-Western Ethiopia: Understanding the Broader Community Risk.

Q fever is a zoonotic disease that is caused by Coxiella burnetii and leads to abortion and infertility in ruminants and debilitating disease in humans. Jimma zone, including Jimma town, located in the Oromia region of Ethiopia, was affected by an outbreak of abortions in ruminants related to Q fever infection between 2013 and 2015. This study aimed to investigate the geo-clustering of C. burnetii seroprevalence in dairy farms of Jimma town and identify the environmental risk factors associated with seroprevalence distribution. A total of 227 cattle were tested for antibodies against C. burnetii in 25 farms. We explored the clustering of C. burnetii seroprevalence using semivariograms. A geostatistical regression-based model was implemented to quantify the risk factors and to predict the geographical variation in C. burnetii seroprevalence at unsampled locations in Jimma town using OpenBugs. Our results demonstrated that the risk of exposure in dairy cattle varied across the landscape of Jimma town and was associated with environmental risk factors. The predictive map of C. burnetii seroprevalence showed that communities in the eastern part of Jimma town had the highest risk of exposure. Our results can inform community-level investigations of human seroprevalence in the high-risk areas to the east of Jimma.

Commensal bacterial sharing does not predict host social associations in kangaroos.

Social network analysis has been postulated as a tool to study potential pathogen transmission in wildlife but is resource-intensive to quantify. Networks based on bacterial genotypes have been proposed as a cost-effective method for estimating social or transmission network based on the assumption that individuals in close contact will share commensal bacteria. However, the use of network analysis to study wild populations requires critical evaluation of the assumptions and parameters these models are founded on. We test (a) whether networks of commensal bacterial sharing are related to hosts' social associations and hence could act as a proxy for estimating transmission networks, (b) how the parameters chosen to define host associations and delineate bacterial genotypes impact inference and (c) whether these relationships change across time. We use stochastic simulations to evaluate how uncertainty in parameter choice affects network structure. We focused on a well-studied population of eastern grey kangaroos (Macropus giganteus), from Sundown National Park, Australia. Using natural markings, each individual was identified and its associations with other kangaroos recorded through direct field observations over 2 years to construct social networks. Faecal samples were collected, Escherichia coli was cultured and genotyped using BOX-PCR, and bacterial networks were constructed. Two individuals were connected in the bacterial network if they shared at least one E. coli genotype. We determined the capacity of bacterial networks to predict the observed social network structure in each year. We found little support for a relationship between social association and dyadic commensal bacterial similarity. Thresholds to determine host associations and similarity cut-off values used to define E. coli genotypes had important ramifications for inferring links between individuals. In fact, we found that inferences can show opposite patterns based on the chosen thresholds. Moreover, no similarity in overall bacterial network structure was detected between years. Although empirical disease transmission data are often unavailable in wildlife populations, both bacterial networks and social networks have limitations in representing the mode of transmission of a pathogen. Our results suggest that caution is needed when designing such studies and interpreting results.

El análisis de redes sociales se ha postulado como una herramienta para estudiar la potencial transmisión de patógenos en fauna silvestre. Sin embargo, las redes sociales en fauna silvestre son difíciles de cuantificar. Es por esto que las redes basadas en el genotipo bacteriano se han propuesto como un método más rentable para estimar redes sociales o de transmisión, basado en el supuesto de que individuos en contacto cercano compartirán bacterias comensales. Sin embargo, el uso del análisis de redes para estudiar poblaciones silvestres requiere una evaluación crítica de los supuestos y parámetros en los que se basan estos métodos. En este estudio probamos (a) si las redes de bacterias comensales están relacionadas con las redes sociales y por lo tanto podrían actuar como proxy para estimar redes de transmisión. Exploramos (b) cómo los parámetros elegidos, para definir asociaciones entre individuos y delinear genotipos bacterianos, impactan en los resultados. Finalmente, examinamos (c) si estas relaciones cambian con el tiempo. Utilizamos simulaciones estocásticas para evaluar cómo la incertidumbre en la elección de parámetros afecta la estructura de la red. Nuestro estudio se basó en una población de canguros grises (Macropus giganteus), del Parque Nacional de Sundown, Australia. Cada individuo fue identificado usando marcas naturales, y sus interacciones con otros canguros se registraron a través de observaciones de campo durante dos años para construir redes sociales. Se recolectaron muestras fecales, se cultivó Escherichia coli, se determinó el genotipo mediante BOX-PCR y se construyeron las redes bacterianas. El criterio para conectar dos individuos en la red bacteriana, se basa en sí al menos un genotipo de E. coli es compartido entre dos individuos. Determinamos la capacidad de las redes bacterianas para predecir la estructura de las redes sociales observadas en cada año. Encontramos poca evidencia que sustente una relación entre la asociación social y la similitud bacteriana. Los umbrales, que determinan las asociaciones entre individuos en las redes sociales y los valores de corte de similitud para definir los genotipos de E. coli, tuvieron una importante influencia para inferir vínculos entre individuos. De hecho, encontramos que las inferencias pueden mostrar patrones opuestos dependiendo de los umbrales elegidos. Además, no se detectó similitud en la estructura de la red bacteriana general entre los años. Aunque los datos empíricos de transmisión de enfermedades a menudo no están disponibles para poblaciones de vida silvestre, los análisis de redes pueden potencialmente solventar dicho problema. Sin embargo, tanto las redes bacterianas como las redes sociales tienen limitaciones para representar el modo de transmisión de un patógeno. Nuestros resultados sugieren que se necesita tener cautela al diseñar este tipo de estudios e interpretar los resultados.

Drivers of Bartonella infection in micromammals and their fleas in a Mediterranean peri-urban area.

People living at the human/wildlife interface are at risk of becoming infected with Bartonella for which micromammals act as reservoir. We aimed to determine the factors related to the prevalence of Bartonella and its haplotype diversity in micromammals and in their fleas in a Mediterranean peri-urban environment. We analyzed 511 micromammals, chiefly 407 wood mice (Apodemus sylvaticus), captured into Barcelona metropolitan area (Spain) in spring and autumn from 2011 to 2013 in two natural and two adjacent residential areas, their fleas (grouped in 218 monospecific pools) and 29 fetuses from six Bartonella-positive female wood mice. Amplification of a fragment of ITS was carried out by real time PCR. Prevalence was 49% (57% in the dominant species, the wood mouse), and 12 haplotypes were detected. In general, prevalence was higher in those hosts more heavily infested by fleas, coincident with higher rates of capture, in autumn than in spring, and in adults than in juveniles. Prevalence did not differ between natural and residential areas except for one prevalent haplotype, which was more frequent in natural areas. Prevalence in flea pools (58%) was only explained by Bartonella occurrence in the pool host. In 56.4% of the flea pools with identified Bartonella haplotypes, we found the same haplotype in the host and in its flea pool. Prevalence in wood mouse fetuses was 69%, with at least one infected fetus in all litters, and two litters with all the fetuses infected. indicating that vertical transmission might be important in Bartonella epidemiology in the wood mouse. There is a hazard of Bartonella infection for people living in residential areas and those visiting peri-urban natural areas in Barcelona.

Differences in the ectoparasite fauna between micromammals captured in natural and adjacent residential areas are better explained by sex and season than by type of habitat.

We compared the ectoparasite fauna in 608 micromammals (chiefly 472 wood mice Apodemus sylvaticus, 63 Algerian mice Mus spretus, and 51 greater white-toothed shrews Crocidura russula) captured in natural and adjacent residential areas in spring and autumn during three consecutive years in four areas in periurban Barcelona (NE Spain). We found little support for an association of urbanization with differences in infestation by ectoparasites. Prevalence of Rhipicephalus sp. tick in wood mice and shrews was significantly higher in residential than in natural habitats, and the opposite was found for the flea Ctenophtalmus andorrensis catalanensis in shrews. Marked differences in the prevalence of the flea Leptopsylla taschenbergi amitina in wood mice between seasons were observed in natural but not in residential habitats, probably due to enhanced flea survival probabilities in the latter. However, as a rule, males were more frequently and heavily infested than females, and the prevalence was higher in autumn than in spring. Our results suggest that the ectoparasite fauna of periurban micromammals is shaped more by other factors than by habitat modification. People living in residential areas are at risk of contact with the arthropods borne by non-commensal micromammals and the pathogens transmitted by them.

Molecular detection of vector-borne pathogens in wild and domestic carnivores and their ticks at the human-wildlife interface.

Urbanization of natural areas is considered one of the causes of the current apparent emergence of infectious diseases. Carnivores are among the species that adapt well to urban and periurban environments, facilitating cross-species disease transmission with domestic dogs and cats, and potentially with their owners. The prevalence of vector-borne pathogens (VBP) of zoonotic and veterinary interest was studied in sympatric wild and domestic carnivores into Barcelona Metropolitan Area (NE Spain). Blood or spleen samples from 130 animals, including 34 common genets (Genetta genetta), 12 red foxes (Vulpes vulpes), 10 stone martens (Martes foina), three Eurasian badgers (Meles meles), 34 free-roaming domestic cats and 37 dogs with outdoor access, were collected either in protected or adjacent residential areas. A total of 309 ticks (chiefly Rhipicephalus turanicus) were collected on these animals. The samples were analyzed with a battery of PCR assays targeting the DNA of Rickettsia spp., Anaplasmataceae, Coxiella burnetii, Bartonella spp., and Piroplasmida, and the amplicons were sequenced. The fox showed the highest prevalence (58%) and diversity of VBP (four pathogens), whereas none of the dogs were infected. Bartonella spp. (including B. clarridgeiae, B. henselae, and B. rochalimae) was the most prevalent pathogen. Infection of wild carnivores with Ehrlichia canis, C. burnetii, Theileria annae and Babesia vogeli was also confirmed, with some cases of coinfection observed. The presence of DNA of T. annae and B. vogeli was also confirmed in tick pools from four species of wild carnivores, supporting their role in piroplasmid life-cycle. By the sequencing of several target genes, DNA of Rickettsia massiliae was confirmed in 17 pools of Rh. turanicus, Rh. sanguineous, and Rh. pusillus from five different species, and Rickettsia conorii in one pool of Rh. sanguineous from a dog. None of the hosts from which these ticks were collected was infected by Rickettsia. Although carnivores may not be reservoir hosts for zoonotic Rickettsia, they can have an important role as mechanical dispersers of infected ticks.

Infection and exposure to vector-borne pathogens in rural dogs and their ticks, Uganda.

BACKGROUND: In rural parts of Africa, dogs live in close association with humans and livestock, roam freely, and usually do not receive prophylactic measures. Thus, they are a source of infectious disease for humans and for wildlife such as protected carnivores. In 2011, an epidemiological study was carried out around three conservation areas in Uganda to detect the presence and determine the prevalence of vector-borne pathogens in rural dogs and associated ticks to evaluate the risk that these pathogens pose to humans and wildlife. METHODS: Serum samples (n = 105), blood smears (n = 43) and blood preserved on FTA cards (n = 38) and ticks (58 monospecific pools of Haemaphysalis leachi and Rhipicephalus praetextatus including 312 ticks from 52 dogs) were collected from dogs. Dog sera were tested by indirect immunofluorescence to detect the presence of antibodies against Rickettsia conorii and Ehrlichia canis. Antibodies against R. conorii were also examined by indirect enzyme immunoassay. Real time PCR for the detection of Rickettsia spp., Anaplasmataceae, Bartonella spp. and Babesia spp. was performed in DNA extracted from FTA cards and ticks. RESULTS: 99% of the dogs were seropositive to Rickettsia spp. and 29.5% to Ehrlichia spp. Molecular analyses revealed that 7.8% of the blood samples were infected with Babesia rossi, and all were negative for Rickettsia spp. and Ehrlichia spp. Ticks were infected with Rickettsia sp. (18.9%), including R. conorii and R. massiliae; Ehrlichia sp. (18.9%), including E. chaffeensis and Anaplasma platys; and B. rossi (1.7%). Bartonella spp. was not detected in any of the blood or tick samples. CONCLUSIONS: This study confirms the presence of previously undetected vector-borne pathogens of humans and animals in East Africa. We recommend that dog owners in rural Uganda be advised to protect their animals against ectoparasites to prevent the transmission of pathogens to humans and wildlife.

Factors associated with the prevalence and pathology of Calodium hepaticum and C. splenaecum in periurban micromammals.

Calodium hepaticum (syn. Capillaria hepatica) and Calodium splenaecum (syn. Capillaria splenaecum) are nematodes that infect the liver and spleen, respectively, of mammals. While the host range, distribution, pathology and zoonotic potential of C. hepaticum are well known, very little is known about C. splenaecum. The observed prevalence of these two parasites, the factors associated with prevalence, and the lesions resulting in the different host species were studied in 408 micromammals captured in two periurban areas of Barcelona (NE Spain) from 2011 to 2013. C. hepaticum was found in 4% of 322 wood mice (Apodemus sylvaticus) (with local prevalence up to 16%) and 1 of 2 Norwegian rats (Rattus norvegicus). C. splenaecum was found in 10 % of 38 greater white-toothed shrew (Crocidura russula) (local prevalence up to 30%). Neither parasite was detected in 29 Algerian mice (Mus spretus) and 17 black rats (Rattus rattus). Prevalence of C. hepaticum was significantly higher in wood mice captured in natural areas (6.4%) than those from residential areas (0%), and infected mice were in better body condition. No differences in prevalence were found among age and sex groups, years and seasons. Lesions of hepatic capillariasis in wood mice consisted mainly of mild to moderate multifocal granulomas around degenerating adult parasites and/or eggs, while lesions seen in a rat consisted of multifocal granulomatous hepatitis and bridging fibrosis extending from the necrotic areas caused by the parasites. Splenic lesions found in shrews due to C. splenaecum, representing the first histological description of this parasite, were single nodules that corresponded to finely encapsulated clusters of eggs with adult parasites.