Seropositivity and Risk Factors Associated with Toxoplasma gondii Infection in Wild Birds from Spain

Toxoplasma gondii is a zoonotic intracellular protozoan parasite of worldwide distribution that infects many species of warm-blooded animals, including birds. To date, there is scant information about the seropositivity of T. gondii and the risk factors associated with T. gondii infection in wild bird populations. In the present study, T. gondii infection was evaluated on sera obtained from 1079 wild birds belonging to 56 species (including Falconiformes (n = 610), Strigiformes (n = 260), Ciconiiformes (n = 156), Gruiformes (n = 21), and other orders (n = 32), from different areas of Spain. Antibodies to T. gondii (modified agglutination test, MAT titer ≥1∶25) were found in 282 (26.1%, IC95%:23.5–28.7) of the 1079 birds. This study constitute the first extensive survey in wild birds species in Spain and reports for the first time T. gondii antibodies in the griffon vulture (Gyps fulvus), short-toed snake-eagle (Circaetus gallicus), Bonelli's eagle (Aquila fasciata), golden eagle (Aquila chrysaetos), bearded vulture (Gypaetus barbatus), osprey (Pandion haliaetus), Montagu's harrier (Circus pygargus), Western marsh-harrier (Circus aeruginosus), peregrine falcon (Falco peregrinus), long-eared owl (Asio otus), common scops owl (Otus scops), Eurasian spoonbill (Platalea leucorodia), white stork (Ciconia ciconia), grey heron (Ardea cinerea), common moorhen (Gallinula chloropus); in the International Union for Conservation of Nature (IUCN) “vulnerable” Spanish imperial eagle (Aquila adalberti), lesser kestrel (Falco naumanni) and great bustard (Otis tarda); and in the IUCN “near threatened” red kite (Milvus milvus). The highest seropositivity by species was observed in the Eurasian eagle owl (Bubo bubo) (68.1%, 98 of 144). The main risk factors associated with T. gondii seropositivity in wild birds were age and diet, with the highest exposure in older animals and in carnivorous wild birds. The results showed that T. gondii infection is widespread and can be at a high level in many wild birds in Spain, most likely related to their feeding behaviour.


Introduction
Toxoplasma gondii is a zoonotic intracellular protozoan parasite of worldwide distribution [1,2]. Wild and domestic felids are the definitive hosts, excreting oocysts in faeces. Humans and virtually all warm-blooded species, including birds, can be intermediate hosts and can become infected by ingestion of food and water contaminated with sporulated T. gondii oocysts, by consumption of tissue cysts in infected animal tissues, or congenitally [1,2].
Toxoplasma gondii infection is prevalent in many avian species including poultry, game and other species in the wild [3,4,5,6,7,8,9]. In addition, clinical cases have been reported and T. gondii considered to be the cause of mortality in birds of different species [3,4,5,10,11]. However, there is still only scant information about seroprevalence of T. gondii and the risk factors associated to T. gondii infection in wild bird populations. Investigation of T. gondii seropositivity in birds could be a useful way to assess environmental contamination with oocysts since some avian populations feed directly on the ground and are continuously exposed to oocyst ingestion [11]. Furthermore, the investigation of T. gondii infection in avian scavengers might allow the assessment of the possibility of contact with intermediate hosts and the associated risk for public health [12].
The present study analyzed T. gondii seropositivity and provides information on possible associated risk factors, i.e. species, order, dietary habits, migration, sex, age and geographic origin in numerous species of wild birds from Spain, including some endangered species listed in the International Union for Conservation of Nature (IUCN) Red List of Threatened Species [13].

Ethics statement
Samples were collected in compliance with the Ethical Principles in Animal Research in the wildlife rehabilitation centres. The study of this material (blood serum) did not require the approval of an ethics committee because the collection of sera is considered among routine procedures and the sera was collected and stored before the design of this study.
Samples were collected from wild birds captured or brought to wildlife rehabilitation centres and other breeding centres between 1996 and 2010. Blood collection was from the jugular vein, the heart, or the brachial vein. Blood was placed in a serum collection tube until clotted and then centrifuged. Sera were stored at 220uC until analysis. The samples were obtained from the main geographic areas from Spain including Northern Spain (n = 332) (Catalonia (228 samples) and Castile and León (4 samples from Valladolid); Central Spain (Castile La Mancha) (n = 565); and Southern Spain (Andalusia) (n = 282). The provinces sampled within Catalonia were Lleida and Barcelona; within Castile-La Mancha the provinces of Toledo and Ciudad Real, and within Andalusia the provinces of Granada, Málaga, Cádiz, Huelva, Seville, Córdoba and Jaén.

Serological test
Sera were examined by the modified agglutination test (MAT) to detect antibodies against T. gondii as described previously [14]. Sera were tested at 1:25, 1:50, 1:100 and 1:500 dilutions. A commercial positive control (Toxotrol-A, Biomerieux, France) diluted from 1:25 to 1:3200 (with a minimum titre of 1:200) was included in each test. Negative controls were also included in all tests. Titres of 1:25 or higher were considered positive and those with doubtful results were re-examined. This technique has been previously evaluated in several bird species [3,15].

Definition of variables and Statistical analysis
Bird species were classified in regard to feeding behaviour as carnivorous (birds of prey), scavenger (when they feed on dead or decaying matter), granivorous-herbivorous-insectivorous (G/H/I), omnivorous and piscivorous. Whenever possible, sex and age data were collected. Age was classified in two categories (,1 year and $1 year). In addition, wild birds were classified as migratory or sedentary (non-migratory), according to their movements [16].
The variables analyzed included wild bird species, order, feeding behavior, geographic area of sample collection, migratory behavior, sex and age. For all statistical analysis only those species with more than 3 analyzed samples (birds) were included.
Association between explanatory variables and T. gondii seropositivity was tested in three steps. Firstly, a bivariate chisquare test was performed to obtain an indication of the relevance of the explanatory variables to the risk of an animal being seropositive. When observations per category were less than six, Fisher's exact test was used. Secondly, factors showing a pvalue,0.25 were further scrutinized for associations using Spearman's rank correlation coefficient (r) to avoid colinearity problems. If r was larger than 0.4 the variable more clearly linked to T. gondii infection, based on epidemiological data and previous studies in the literature, was retained. The third step involved a multiple logistic-regression model using a non-automatic backward selection of variables. Biologically plausible confounding factors were tested using Mantel-Haenszel analysis and confounding was considered to be potentially significant if ORs shifted appreciably. Changes in the OR greater than 30% were considered indicative of confounding. The fit of the models was assessed using the Hosmer and Lemeshow goodness-of-fit test [17]. Potential twoway interactions between the variables were tested for significance in the model. The model was re-run until all the remaining variables presented statistically significant values (likelihood-ratio Wald's test, p,0.05), and a potential causal relationship with the response variable existed. Statistical analyses were performed using SPSS 15.0 (Statistical Package for Social Sciences (SPSS) Inc., Chicago, IL, USA).
In the bivariate analysis, considering those species with more than 3 samples analyzed, statistically significant differences were observed between seropositivity levels and the variables: order, wild bird species, geographic area of sample collection, migratory behaviour, age, year of sampling and feeding behaviour.
Based on their feeding behaviour, carnivorous wild birds showed statistically significant higher seropositivity of T. gondii (36.2% of 588) compared to scavenger (15.6% of 262 samples), The order Strigiformes showed statistically higher seropositivity (44.0% of 259) compared to all the other orders (P,0.01). In addition, the seropositivity in the order Falconiformes (23.8% of 602) was significantly higher than that in the order Ciconiiformes (11% of 154) (P,0.05). Seropositivity in Gruiformes order was 5.6% (1 of 18), and in other orders was 12.5% (2 of 16).
The positivity of infection was significantly higher (P,0.001) in samples collected from wild birds from Northern (mostly Catalonia) and Central Spain (27.2 and 30.2%, respectively) compared to Southern Spain (18.3%), (P,0.001).
Significant differences were observed between samples from birds older than 1 year showed statistically significant higher seropositivity (38.6% of 158) than birds younger than 1 year (9.5% of 148). Significant differences were also observed also among years of sampling (P,0.001). Samples collected up to the year 2000 showed high seropositivity but the number of analyzed The final multivariate model showed that the main risk factors associated with T. gondii seropositivity in wild birds were age and feeding behaviour. Older wild birds had a 6.8 times higher risk of    Table 2).

Discussion
The present study included a wide range of wild bird species and a high number of analyzed samples supplements the existing data on seropositivity of T. gondii in wild birds worldwide, and adds numerous new species of wild birds to the possible intermediate host list for T. gondii infection. To our knowledge, this is the first report of T. gondii antibodies in griffon vulture, short-toed snakeeagle, Bonelli's eagle, golden eagle, bearded vulture, osprey, Montagu's harrier, Western marsh-harrier, peregrine falcon, longeared owl, common scops owl, Eurasian spoonbill, white stork, grey heron, common moorhen, the IUCN ''vulnerable'' Spanish imperial eagle, lesser kestrel and great bustard; and the IUCN ''near threatened'' red kite and it is the first survey of these wild bird species in Spain. Detection and/or isolation of the parasite from these species would be necessary to corroborate that these species are intermediate hosts of T. gondii.
The results showed that T. gondii infection is widespread in many wild birds in Spain, with high variation among different species, orders, geographical regions and feeding behaviour. The main risk factors associated with T. gondii seropositivity in wild birds were age and feeding behaviour, with higher exposure observed in older animals and in species with a meat-based diet. The observed age effect is in accordance with previous studies of T. gondii seroprevalence in other animal species such as pigs, Iberian lynx (Lynx pardinus), Spanish ibex (Capra pyrenaica) and cats in Spain [18,19,20,21] as well as in numerous studies of other species worldwide [2], and would be an indication of cumulative likelihood for exposure to T. gondii with age [2,22] and to lifelong persistence of antibodies. The higher seropositivity observed in older individuals could also be an indication of horizontal transmission as the main route of transmission of T. gondii in these species.
Carnivorous species showed a significantly higher seropositivity of T. gondii. In this respect, the order Strigiformes (in which virtually all the species are carnivorous) showed the highest seropositivity, followed by the order Falconiformes, in which the majority of species are carnivorous and/or scavenger species. The order Gruiformes, with mostly G/H/I species, showed the lowest seropositivity. This agrees with the results obtained in mammals, in which several studies have observed that carnivorous species have higher seroprevalence of T. gondii antibodies compared to omnivorous, herbivorous or insectivorous species [23,24,25,26]. A possible explanation could be the cumulative ingestion of infected prey, as suggested by similar results from comparison of seroprevalence levels in carnivorous mammals such as cats [21], omnivorous such as wild boars [27] and herbivorous such as red deer and other wild ruminants [28] in Spain. The main mode of infection of wild birds, as of other intermediate hosts, with T. gondii is through ingestion of food and water contaminated with sporulated oocysts (mainly in ground feeding birds) and by consumption of cysts contained in tissues of infected animals (mainly in carnivorous, scavenger birds) [2]. The present results suggest that the main mode of transmission of T. gondii in the analyzed wild birds was consumption of tissue cysts.
In agreement with the above results, the highest seropositivity was observed in the Eurasian eagle owl (68.1% of 144), a carnivorous species recently included in the list of birds with T. gondii antibodies by Lopes et al [6]. High seropositivity was also observed in other carnivorous species such as the common buzzard, which has been previously reported to have high seroprevalence in France (79% of 14 samples) [4], and in the golden eagle. The main diet of these species of carnivorous wild birds includes rodents and other small mammals, which are well known reservoirs of T. gondii infection for many animal species, especially pigs [25,29,30]. On the other hand, high seropositivity was also observed in some omnivorous species, such as the grey heron (60% of 5) and in some species of duck [Anas acuta (100% of 1) and A. plathyrhynchos (33.3% of 6)], although the number of analyzed samples was relatively low. The grey heron mainly feeds in shallow water catching fish, frogs, and insects, but will also eat small mammals, reptiles and other birds [31] and, like other Ciconiiformes, has been moving into urban environments where they also make use of food discarded by humans. The fact that grey heron samples were mainly from Catalonia (4 of 5 samples, with all the positive samples from this location), which seems to be one of the most prevalent areas of T. gondii infection in Spain, with higher prevalence of T. gondii compared to other locations for species such as wild rabbits [32], roe deer [33] and domestic pigs [18], and where higher detection of the parasite has also been observed in tissues from wild birds [9], could also be an explanation for the increased seropositivity observed in this species. Moderate precipitation and humidity in those areas might favour oocyst survival and sporulation in the environment facilitating T. gondii spread and maintenance [1]. The ducks analyzed in the present study were from different geographical regions, indicating that seroprevalence was not due to location. High seroprevalence in ducks has been recently reported by Alvarado-Esquivel et al [8] (2 positive samples, in 1 A. platyrhynchos and 1 A. diazi, of 4 ducks analyzed). Similarly, it was also unexpected to observe high seropositivity of T. gondii in bearded vultures (G. barbatus) (42.9% of 15) because this species feeds primarily upon livestock bones. A possible explanation is that during the chick-rearing period, small and medium-sized mammals and other birds are also included in the chicks, and even in the parents diet [34]. Furthermore, most bearded vultures were also obtained in the northern area. Further studies including more samples of this species will need to be performed to confirm this observation.
To the contrary, the Egyptian vulture, a small vulture inhabiting open landscapes in arid and rugged regions in the Mediterranean, Africa and Asia, did not show antibodies to T. gondii. The diet of Egyptian vultures is based on carcasses of small and medium-sized wild and domestic animals and, in a lesser degree, livestock, usually ingesting small pieces of skin or muscle [35] which could be the reason for the observed lack of contact with T. gondii in the analyzed samples.
Interestingly, some piscivorous birds showed contact with the parasite, including some species such as the Eurasian spoonbill (P. leucorodia) analyzed in high numbers, (6.2% of 81 samples). This result could be an indication of water contamination by oocysts. Until recently, waterborne transmission of T. gondii was considered uncommon. However, a large human outbreak linked to contamination of a municipal water reservoir in Canada by wild felids and the widespread infection of marine mammals in the USA provided reasons to question this view [36]. Toxoplasma gondii antibodies have been observed in Mediterranean wild dolphins [37] and a case of systemic toxoplasmosis in a pregnant dolphin from the Spanish Mediterranean coast has been reported [38].
In summary, the present study extends the range of species of wild birds that have antibodies to T. gondii and indicates widespread exposure to T. gondii among many wild bird species in Spain, with highly variable postitivity levels among species and orders. The main risk factors associated with T. gondii seropositivity in wild birds were age and feeding behaviour, with higher exposure in older animals and in carnivorous wild birds. The results indicate that T. gondii infection can be high in some species, mainly related to their feeding behaviour, and that these species may have an important role in the epidemiology of T. gondii infection. The observed contact with the parasite in some vulnerable or near threatened wild bird species suggests that the analysis of this parasite should be taken into account in relevant wild bird conservation programs.