Epidemiology of Q fever in domestic ruminants in the central zone of the Iberian peninsula
PhD Thesis defense by María Teresa García-Seco Romero at the VISAVET Centre of the Complutense University of Madrid
June 22nd, 2017
Coxiella burnetii is the aetiological agent of Q fever, a worldwide distributed zoonosis whose main reservoirs and source of infection for humans are domestic ruminants. The main clinical sign associated with C. burnetii infection is late-term abortion in small ruminants, and fertility problems in cattle. However, the impact of Q fever in animal production has been traditionally considered limited since infection is more commonly unapparent. Still, its ability to cause important clinical outbreaks has become evident in the last years, so that nowadays it is considered one of the main causes of abortion in domestic ruminants. Infected animals may excrete high amounts of bacteria during the abortion/parturition, contaminating the environment and potentially leading to new infections in susceptible individuals. The control of C. burnetii is difficult due to its low infective dose, its high environmental resistance and the potential for airborne transmission events over long distances when conditions are favorable. In fact, airborne transmission due to inhalation of contaminated aerosols is considered the main route of infection in humans.
In humans, C. burnetii infection is also normally silent (in about 60% of the cases). However, in a proportion of causes it may cause acute and chronic infections that, in occasions, may be highly incapacitating. Moreover, in certain cases the complication of clinical signs may be fatal. These characteristics, together with the increasing number of reported cases during the last decade (both sporadic as well as part of outbreaks affecting thousands of people, generally associated to domestic ruminants) have revealed the potential clinical consequences this traditionally neglected disease can have in public health. Nevertheless, there are numerous knowledge gaps in several relevant clinical and epidemiological aspects of the disease.
Thus, in the last years there has been a growing interest in assessing different aspects of Q fever, leading to an increase in research and surveillance activities by public agencies and international organizations. However, although Q fever is current a disease subjected to mandatory reporting in many countries, in most of them there are no official surveillance programs, what results in a lack of knowledge about the real situation of Q fever in domestic ruminants in many regions where the prevalence of the disease is considered subestimated.
Several regions in Spain, such as the Basque Country or the Canary Islands, have reported multiple human cases in the last decades, sometimes as outbreaks. Because of this there has been a stronger focus on the study of Q fever in domestic ruminants in these areas. However, there are many other regions in Spain where the information about Q fever in domestic ruminants is lacking in spite of the multiple human cases notified in them.
In this context, the objective of the present work was to determine the epidemiological situation of Q fever in domestic ruminants in the central region of Spain - Castilla y León and Madrid regions –, areas with an important livestock activity but in which the information about Q fever was almost inexistent.
In order to achieve this objective three studies were performed: study 1, focused on sheep from Castilla y León from 2009 to 2012; study 2, in cattle from Madrid in 2009; and study 3, in small ruminants from Madrid in 2011.
Sheep included in the study 1 were selected among those included in the framework of a tularemia surveillance program in lagomorphs and other species (“Procedimiento Integral de Investigación Epidemiológica de Tularemia y otras Zoonosis Asociadas a Roedores Silvestres y Lagomorfos de la Comunidad Autónoma de Castilla y León”) in 2009-2012. Animals included in studies 2 and 3 were selected among those sampled in the framework of the National Programs of eradication of bovine and small ruminant brucellosis, respectively.
All three studies were based on the analysis serum samples to determine the level of exposure to C. burnetii in the livestock populations. Samples were tested using a commercial ELISA kit, allowing to estimate the seroprevalence of C. burnetii infection in the selected species and regions of study. In addition, we collected information about the location of the farms and the characteristics of the animals (studies 2 and 3) and farms (studies 1, 2 and 3) sampled, as well as of counties from which the farms originated (study 1).
The results of the three studies demonstrated that exposure to C. burnetii in all domestic ruminant populations evaluated was widespread. In study 1 we observed during the 2009-2012 period prevalences around 50% and 10% at the herd and individual level, respectively, although a particularly high value was detected in 2012 (76% and 21% at the herd and individual level).
In studies 2 and 3 the herd-level estimated prevalence rates were around 30% in cattle and 47% in small ruminant. Prevalence values at the individual level were 7% in cattle, 14% in sheep and 22% in goats, suggesting a higher level of exposure to C. burnetii in small ruminants than in cattle.
The serological results were analyzed in combination with the data on the geographic location of the farms to evaluate the spatial distribution of Q fever in sheep in Castilla y León (study 1) and cattle (study 2), sheep and goats (study 3) in Madrid. In addition, the possible presence of clustering at the global and local levels was also assessed.
In study 1 we observed that the distribution of positive sheep and flocks at the county level was indicative of global and local spatial clustering during the four years of the study, so that neighboring counties of a positive county had a higher probability of testing positive (up to the fourth level of neighborhood). Also, a high incidence cluster located in the same region was identified in all four years.
The spatial pattern observed in Madrid was less clear than in Castilla y León. The presence of global clustering in the distribution of positive cattle herds was identified up to the first level of neighborhood, but no evidence of global clustering was found for small ruminants. Similarly, no significant local clusters were found for both all cattle and small ruminant herds. This fact could be due in part to the high percentage of positive farms in the Madrid region, what could hamper the ability of the tests to identify spatial patterns.
Finally, a risk factor analysis was conducted in the three studies by looking at county, herd and individual characteristics potentially associated with the results obtained in the serological analysis. A Bayesian approach was applied in study 1 while frequentist multivariable models were fitted for studies 2 and 3.
In study 1, the risk of seropositivity in sheep of Castilla y León was associated, depending on the year, to the productive type (2009-2011), flock size (2009-2011) and intensive management practices (2011 and 2012). Thus, a higher risk was identified in sheep from dairy flocks, under an intensive/mixed management and with a higher census, which is in accordance with other studies on Q fever performed in domestic ruminants. This analysis also allowed us to identify a spatial pattern in the distribution of positive cases, which was only partially explained by the density of small ruminants from the county of origin of the farm.
In studies 2 and 3 the risk of seropositivity in cattle and small ruminant farms from Madrid was associated, with the productive type, suggesting a higher risk in dairy farms as also found in Castilla y León. In addition, larger census was also identified as a significant risk factor for bovine herds. Regarding the individual variables, only the age of the animals was found to be significantly associated with the serological status, so that older animals had an increased risk of testing positive in the ELISA assay, what could be related with a longer chance of exposure to C. burnetii. This risk factor has also been previously reported in other studies.
Overall our results demonstrate that exposure to C. burnetii in the domestic ruminants’ species studied in Castilla y León and Madrid is widespread. In addition, a spatially uneven distribution of the risk of seropositivity was observed, particularly in the case of Castilla and Leon. This information, together with the risk factors identified in each of the areas of study, could be useful to design surveillance programs for Q fever with the purpose of controlling and decreasing the exposure to the bacteria as well as preventing potential transmiss
