EP is widely present in many countries of the world, so much so Schnittger et al. (2012)  estimated that around 90% of the horses live in endemic areas. Thus EP should always to be considered in the differential diagnosis, together with other diseases such as leptospirosis, equine infectious anaemia and anaplasmosis, in case of any anaemic/haemolytic clinical forms, as presence of anaemia, jaundice or petechiae. This study was set with the aim of investigating if individual and management characteristics could influence the probability of being infected. Furthermore it was assessed the possible association of some haemato-biochemical alterations with confirmed cases of T. equi infection, comparing them to healthy subjects and to those ones with compatible signs but not confirmed as T. equi infected. The original purpose was to investigate both EP agents but, since only one case of B. caballi was detected among the subjects eligible, the study was focused on T. equi.
Symptoms are not often reported in details in the studies on EP. Comparing the present study with the few others for which clinical data are reported or easily computable (Fig. 1) a variability in the presence of symptoms emerges, with the predominance (presence in ≥ 50% of confirmed cases) of anaemia, jaundice, lack of appetite, lethargy, petechial haemorrhages, weakness and weight loss.
Gender and species have no influence on the presence of T. equi, while age is a risk factor for T. equi and other diseases, with a cumulative age-dependent increasing presence of the pathogen. Positive cases increase in the groups 7–12 and > 13 years of age compared to the 0–6 years group. This result was expected and in agreement with previous studies , even though it disagrees both with the results of a study by the authors of the same regions , reporting that OR decreased with age, and with Steinman et al. (2012)  that, in Israel, observed no significant differences within the age classes considered. This evidence could be due to the cumulative risk of being in contact with an infected ticks through the life, to the inefficacy of the treatment, that can be variable depending on the strain involved  and to the fact that T. equi is not self-cleared as B. caballi but leads to a chronic infection status.
Breed analysis highlighted a lower risk (OR = 0.08, [95% CI 0.01–0.95]) to be infected in the galopper breed compared to others. It could be a proxy related to different management practices, according to breed and attitude, i.e. more appropriate rearing conditions of pure breed horses and of those ones with economical value, that resulted less exposed to ticks [25, 27]. Because of the sampling was carried out almost only amongst animals belonging to saddler breed and with equestrian attitude, despite authors recommendations to practitioners, no conclusions can be certainly inferred.
The assessment of the fur brightness was included in the study as well as previously considered by Katiyatiya et al. (2015)  in the bovine, assuming that dark colour furs could result in a higher tick load and, consequentely, in an increased risk of infection. However, the present survey is still at a preliminary stage and further studies will be needed to clarify a possible association between T. equi infection and fur colours.
T. equi confirmed cases were approximately four times more likely to be found amongst animals having an external access compared to controls, reinforcing the evidence that access to the outside, including not only a paddock near the stable but also pastures and free-ranging animals, in which the infection risk could be higher for a greater exposure to ticks or for a lack of grooming practices that improve tick removal, is a significant variable, in agreement with previous studies [25, 29, 30].
Vaccinations against other equine diseases and ectoparasiticide treatment seems to have a protective effect on T. equi infection. While for the latter variable the association with a lower presence of infection is expectable, the correlation with vaccination is not clear; it is opinion of the authors that this evidence could be explained with prophylactic treatments being a proxy for good management practices that prevent the infection.
Several studies on haemato-biochemical parameters associated with EP are available for both horses and donkeys. In the present study, the authors performed a comparison of haemato-biochemical parameters among the three groups. AST value resulted lower in the case group, even if not statistically significant (p = 0.06). This finding differs from previous studies [20, 22, 31, 32] reporting an increase, even if not constant, of this enzyme in acute cases as well as for muscle enzyme CK. Both total and direct bilirubin values resulted higher in cases compared to controls, as well as reported by several other studies [19–22, 31], whereas, in the unconfirmed suspects, bilirubin values resulted between those of the other two groups. However only direct birilubin values were over the reference value. A decrease in RBC count, HGB and HCT confirmed T. equi associated findings, caused by the haemolysis of the red blood cells [19–24, 31, 25–28]. The decrease in platelet count is widely reported [20, 21, 31–35], and it is the only parameter that, according to the present data, differs between both cases and controls rather than between cases and non confirmed suspect (Fig. 3). This statistical evidence though, seems to lack any biological explanation justifying it, if compared to other similar disease as B. caballi. In other species, such as dogs with babesiosis, the platelet count decrease is supposed to be caused by sequestration in the spleen, systemic disseminated intravascular coagulation, or immune-mediated depletion . An increase of MCH is also reported in donkeys [19, 33] and horses  while variation in MCV, as an increased value, is reported only by Davitkov et al. (2017)  and Mahmoud et al. (2016)  and as a decreased value only in two out of 16 cases by Pasolini et al. (2018) . In the present study, MCV resulted different only between controls and unconfirmed suspects. MCHC is reported to vary in affected donkeys [19, 33] whereas in the present study resulted not significantly different, as well as for others [20–22, 31, 34, 35]. An increase of WBC count was not observed in the present study despite being reported in several others [18, 20–22, 31, 33, 34] as well as a decrease in total protein [20, 22]. This is the first report describing a decrease in eosinophils, basophils and RDW associated to T. equi infection, in contrast with Mahmoud et al. (2016)  but in agreement with previous studies in cattles with theileriosis . Hypoalbuminemia and hyperglobulinemia were observed with a consequent decrease of the ratio of the two parameters. Usually albumin decreases as a result of liver dysfunction due to the disease. A decrease in albumin was reported also in dogs with babesiosis , but this is the first finding in infected equids, following Zobba et al. (2008)  that described, without inferential analysis, an albumin decrease in 15% of the positive subjects whereas globulin was reported as normal in all of them; and Pasolini et al (2018)  that reported a decreased value in two of 16 positive horses. Finally it has to be highlighted that some of the haemato-biochemical parameters, in particular the total bilirubin, CK, percentage of monocytes, eosynophils, basophils, MCH, RDW, albumin, and gamma fraction, despite being associated to T. equi, are still within the reference values. The same observation has been reported in most of the studies previously mentioned in the discussion, meaning that, in evaluating the laboratory results, these parameters are not useful for practioners, that have to refer only to meaningful parameters such as direct bilirubin, RBC, HGB, HCT and A/G.
Limitations of this study has been already partially discussed in ‘material and methods’ section, as they influence the case definition and data analysis. Interviewer bias brought the major constraint to the study, in fact, in many cases, symptoms reported by the practitioners were not confirmed by the laboratory results, suggesting that the data provided were manipulated to result eligible for the study. The reason could be the following: the Istituto Zooprofilattico Sperimentale del Lazio e della Toscana “M. Aleandri” offered free analysis for the subject enrolled in the study to incentive the collaboration of the equine practitioners, and some of them could have taken advantage of this offer, submitting forms reporting non-observed signs. The double check between signs and laboratory findings aimed to overcome this limitation, excluding subjects with insubstantial data.
The aim of the study, as already mentioned, was to provide valid tools for the clinical activities to the practitioners even if, after discussing the statistical and the construct validity, external validity should be discussed, i.e. whether the study results can be generalized to different settings and times , as other species, breeds and attitudes, and different level of prevalence. Moreover, a similar investigation should be performed also for B. caballi and other diseases in the differential diagnosis of anemic/haemolotic clinical form to provide more data in order to assess a more precise diagnosis.