Zoonotic leishmaniasis and heartworm disease in Italy: changing epidemiological scenarios and future perspectives

Background: For long time, canine leishmaniosis (CanL) was considered endemic in the southern, central, and insular regions of Italy, whereas heartworm disease (HW) by Dirofilaria immitis in the northern region and in the swampy Po valley. Following the reports of new foci of both diseases, in this study we update the distribution patterns and occurrence of new foci of CanL and HW discussing the main drivers for the changes in the epidemiology of these two important zoonotic canine vector-borne diseases. Methods: Based on the statistical analyses of serological assays (n=90,633) on CanL and HW performed by reference diagnostic centres in Italy over a ten-year period (2009–2019), the distribution patterns of both diseases were herein presented along with the occurrence of new foci. Results: Results highlighted the changing distribution patterns of CanL vs HW disease in Italy. CanL is now also endemic in the northern regions and HW has endemic foci in central and southern regions and islands. Significant differences in CanL and HW prevalence among the study macroareas were detected. The overall prevalence of CanL was 28.2% in Southern Italy and Islands, 29.6% in Central Italy and 21.6% in Northern Italy and that of HW of 2.83% in northern Italy, 7.75% in central Italy and 4.97% in southern Italy and islands. HW prevalence significantly varied over years ( χ 2 =108.401, d.f. =10, p <0.0001), gradually increasing from 0.77% in 2009 to values ranging from 5.19-8.47% in 2016-2017. Conclusions: The new epidemiological scenarios have been discussed according to a range of factors (e.g. environmental modifications, occurrence of competent insect vectors, transportation of infected animals to non-endemic areas, chemoprophylaxis or vector preventative measures), which may affect the current distribution. Overall, results advocate for epidemiological surveillance programs, more focussed preventative and control measures even in areas where few or none cases of both diseases have been diagnosed.


Background
In the last decades, canine vector-borne diseases (CVBDs) have been expanding worldwide due to several factors linked with increase in pet travelling along with owners, relocation of stray animals 3 from endemic to previously non-endemic regions, as well as to the modification of the ecology of arthropod vectors and, importantly, environmental modifications [1][2][3]. In some cases, this resulted in the spreading of new parasites and pathogens, and related infections, in previously non-endemic geographical areas, which poses major concerns to veterinary practitioners and, in the case of zoonotic ones, to public health officials [2]. Protozoan Leishmania infantum Nicolle and nematode Dirofilaria immitis Leidy represent paradigmatic examples of the modification in the distribution of the CVBDs they cause (i.e., canine leishmaniosis, CanL, and heartworm disease, HW). The expansion of the above infections has been related to the distribution of their vectors (i.e., for CanL, sandflies of the genus Phlebotomus in the Mediterranean region; for HW, several mosquito species, belonging to the genera Aedes, Anopheles and Culex; [4][5][6][7]).
In specific geographical contexts, such as Italy, where both CVBDs have been endemic for long time [8], their ecology and distribution have been studied and new foci were reported by a retrospective analysis focusing on the period from 1990 to 2009 [1]. Indeed, until 1990, CanL was considered endemic in the southern, central, and insular regions of Italy, whereas dirofilariosis in the northern region and in the Po valley [9,10]. Sporadic case reports suggested that the distribution of both CVBDs has been changing, in that, dirofilariosis expanded towards the southern regions and CanL to the northern regions [1]. This phenomenon could have been the effect of chemoprophylaxis measures by using macrocyclic lactones (ML) in endemic areas of northern Italy and of the absence of such preventative measures in remaining central and southern regions [11]. Indeed, the only species of filarioids historically diagnosed in southern Italy have been Acanthocheilonema reconditum and Dirofilaria repens, causing less pathogenic subcutaneous filariosis [12,13].
Meanwhile, the perception of clinicians and parasitologists has most likely changed, resulting into an increase of the request for diagnostic tests of D. immitis infection in central-southern Italy, as well of CanL in northern regions [8]. In this scenario, the aim of the present study is to present results of serological assays performed by reference diagnostic centres in Italy over a ten-year period (2009 -2019), therefore updating the distribution patterns and occurrence of new foci of CanL and HW.

Databases of serological tests in dogs from Italy
Databases from two diagnostic reference centres (hereafter reported as database A and B) were analysed (total number of serum samples =90,633). Italian macroareas were defined as northern, central and southern/islands macroareas (Fig. 1).
The complied data was standardized to observe significant variations on positive serological results and records of new non-endemic foci. The database A included 62,928 records of dogs collected in a

Statistical analyses
Differences in the prevalence of CanL and HW infection in the studied dogs from northern, central and southern Italy over time were analysed by JMP 9 (SAS) by using weighted generalized linear models (GLZ) with a binomial distribution to test model positive and negative serological outcomes. For each parasitosis, a GLZ with two fixed factors was used to assess significant differences in CanL or HW prevalence among the study macroareas over years: y = Xß + ε where y is the vector of the observation [i.e. serological outcome: positive=1, negative=0, X is the incidence matrix, ß is the vector of fixed effects (the study macroarea (northern, central or southern Italy and islands) and years)] and ε is the vector of the random residual effects (p=0.05).
Then, a dataset was created for each study macroarea and a GLZ with two fixed factors was used to evaluate significant differences in CanL or HW prevalence among the region and years within a given macroarea; the structure of the GLZ was identical to the above described one, with two fixed effects [i.e. the study region and year]. A p-value of 0.05 was used as threshold to assess significant differences among values. To verify that the changing distribution patterns of CanL and HW were not random, a contingency analysis assessing the relationship between the CanL and HW prevalence in the various macroareas, regions and study years was also carried out [14].
Serological data were presented in terms of bi-annual and cumulative prevalence; distribution maps for CanL and HW were generated using QGIS version 3.4.4-Madeira, to compare new non-endemic localities with those reported previously [15].

Results
Overall, the number of serological tests for CanL (n=80,309) performed in the three areas of Italy is reported in Table 1, being higher in southern Italy (37.7% of all tests performed) than in central Italy (31%), and northern Italy (24.7%). Conversely, the overall number of tests requested for the diagnosis of HW throughout Italy (n=10,324) was significantly lower (11.3% of all the requested serological tests) than for CanL, with a relative high number in northern Italy (51.7% of all tests performed) and the lowest in southern Italy (Table 1).
CanL prevalence in this macroarea also showed significant differences among regions (χ 2 =642.949, d.f.=6, p<0.0001) with the highest values recorded in Molise (54.26%), Sicily (50.18%) and Sardinia 7 (38.34%). On the other hand, the study region did not play a significant role impacting HW prevalence, even if a trend was observed (χ 2 =10.723, d.f.=6, p=0.10, respectively). Indeed, the largest number of analysed samples was from Apulia (n = 1608) followed by Basilicata (n = 80) and Sardinia (n = 70). Considering regions with sample size >40, the highest prevalence rates were from Sardinia and Apulia (10%, and 4.73%, respectively). The effect of the study year was not significant  Serological CanL results had a similar cumulative prevalence throughout the Italian territory (Fig. 2 In addition provinces with >300 positive samples were in northern regions (i.e., Turin, Piedmont, and Vicenza Veneto) (Fig. 4).
The highest prevalence of HW was registered in the central regions (i.e., Tuscany and Lazio) followed by southern and northern regions with annual variation patterns throughout the observation period ( Fig. 3). An overall lower number of samples was positive for HW compared to CanL, with >10 positive cases recorded in three provinces from the northern Italy (i.e., Turin, Piedmont; Genova, Liguria; and Belluno, Veneto). The province with the highest number of HW cases in central Italy (i.e., Florence,  [16,18,19] supports the evidence of CanL endemicity and the occurrence of new foci. In addition, clinicians' perception and awareness of the presence of CanL in non-endemic areas seem to be increased in the northern regions, with more than 21,545 tests performed (Table 1). Although the South remains the area with the largest number of tests requested, the large number of tests requested in the North and the Centre indicates that there is a clear awareness of the distribution of the disease throughout Italy.
On the other hand, the cumulative prevalence of HW greatly increased in central (7.7%) and southern Italy and Islands (5%), being higher than in the North (2.8%), which was historically considered the sole endemic area [1,20]. The decreased prevalence of HW in northern Italy regions could be a consequence of clinicians' awareness of the disease and thus of the continuous usage of chemoprophylactic programs in this area. Nonetheless, resistance to ML has been demonstrated in D.
immitis populations from this macroarea [21], wich may represent a potential issue (not yet proven) for re-emergence of resistent strains. Moreover, the low number of tests performed for HW (10.5% of all the requested serological tests) throughout the Italian peninsula also indicates the scant awareness of the occurrence of the disease, especially in southern regions (Table 1). Given the increased prevalence in the Centre, the South and Islands, clinicians should consider the occurrence of HW cases in non-endemic areas.
Furthermore, our analyses showed that the distribution patterns of CanL and HW in Italy are related to a significant relationship between CanL/HW prevalence and the geographical provenience (i.e., macroarea) and study years (2009-2019). Conversely, a major hindrance of this study could be that the anamnesis of animals was not available and therefore their travelling history from/to historically endemic areas for both infections (i.e., northern Italy for HW and southern and central Italy for CanL) performed in that area [19]. In addition, other studies reported cases of CanL and human visceral leishmaniasis diagnosed in the same area, along with their vectors, P. perniciosus and P. neglectus [25].
The southward changing pattern of HW has been detailed in previous researches [26][27][28][29]. A recent questionnaire study pointed out Brindisi and Lecce, from Apulia region, to be areas with more than 20 cases of HW caused by D. immitis [30]. The low prevalence of HW in Sicily and Sardinia could be explained by the small number of test requested or no tests performed in many provinces from these islands. Indeed, earlier studies illustrated the islands have a large number of diagnosed HW [30] as they are suitable for Dirofilaria spp. to thrive [13]. Specifically, Sardinia has the environmental, climatic and human activities (e.g., tourism with animal transportation) that could allow these nematodes to spread [5,30]. Finally, central Italy showed a distribution pattern on which both pathogens are highly prevalent. In this macroarea, the prevalence of CanL was higher (29.6%) of that

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.   Figure 1 Italy. Three main areas with their respective administrative regions.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.