Seroprevalence of enzootic bovine leukosis ( EBL ) : a systematic review and meta-analysis

Enzootic bovine leukosis (EBL) is a viral disease with negative impact on the economy, and the virus has been linked to breast cancer in women and its DNA detected in fresh milk and raw beef for human consumption. In this context, epidemiological surveys allow the knowledge of epidemiological indicators of infection, guiding control programs and consequently making it possible to control and/or eliminate the virus in cattle populations. Therefore, the objective of the present study was to carry out a systematic review with metaanalysis on EBL seroprevalence. Complete articles and brief communications from indexed journals that contained data on the seroprevalence of EBL and described the diagnostic methods used to identify the infection were selected. The study followed the recommendations of the PRISMA methodology Preferred Reporting Items for Systematic Reviews and Meta-Analyzes. In turn, from a total of 581 studies 15 met the eligibility criteria, and the meta-analysis showed a combined prevalence of 31% (95% CI = 25 – 37%), Research, Society and Development, v. 9, n. 12, e43591211228, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i12.11228 3 although with high heterogeneity among the studies, which was attributed to research designs, years of publication of the studies, quantity and age of the animals sampled, exclusive use of dairy herds and heterogeneity among countries. Therefore, this scenario suggests the need for standardization of researches related to epidemiological studies for EBL, specifically crosssectional surveys, with the use of planned sampling, adjustment of rates according to parameters that may influence the prevalence and specific analyzes that provide the determination of reliable epidemiological indicators.

The EBLV can be carried by free particles and by the transfer of cells carrying genetic material. Because it is unstable in the environment, transmission occurs directly between animals or through newly contaminated materials, through body fluids that contain blood or exudate since they are inoculated into the host or in contact with mucous membranes (Hirsch & Leite, 2016). In this context, transmission can occur by rectal palpation, immunization, blood transfusion and surgeries with the use of non-sterilizated materials, milk ingestion, iatrogenic transmission and bites of hematophagous arthropods. Mismanagement practices and deficient hygienic conditions can act as risk factors (Kohara, Takeuchi, Hirano, Sakurai, & Takahashi, 2018;Konishi, Ishizaki, Kameyama, Murakami, & Yamamoto, 2018;Ruiz, Porta, Lomónaco, Trono, & Alvarez, 2018;Panei et al., 2019). Vertical transmission has been proven through the detection of antibodies in newborn calves without previous ingestion of colostrum, and this transmission route is responsible for 3 to 20% of infections in calves born to positive cows (Leuzzi Junior, Alfieri, & Alfieri, 2001).
The diagnosis of EBL can be performed using direct and indirect methods. Direct methods include viral isolation by in vitro culture, and molecular detection of the agent by polymerase chain reaction (PCR), which are very efficient techniques for detecting the agent, however, they are expensive and require qualified inputs and labor. The main indirect methods are the agar gel immunodiffusion test (AGID), officially recognized as the gold standard test, presenting good specificity and sensitivity, relatively simple and low cost, and the enzyme immunoabsorbent assay (ELISA), which presents the best sensitivity and good specificity, ability to detect antibodies in serum and milk, and enable the testing of a high number of animals, presenting as a disadvantage the high cost with equipments for reading and interpretation and the import of diagnostic kits (Hirsch & Leite, 2016;OIE, 2018).
The occurrence of the disease has a negative impact on the economy, since the infection can cause reduced fertility, decreased milk production, increased costs with replacement of heifers, loss of income resulting from the premature slaughter of animals and commercial restrictions (Bartlett et al., 2013;Khudhair, Hasso, Yaseen, & Al-Shammari, 2016;Norby, Bartlett, Byrem, & Erskine, 2016). It's worthy mentioning that BLV has been linked to breast cance in women (Baltzell et al., 2018;Schwingel, Andreolla, Erpen, Frandoloso, & Kreutz, 2019) and BLV DNA detected in fresh milk and raw beef for human consumption (Olaya-Galán et al., 2017). Research, Society and Development, v. 9, n. 12, e43591211228, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i12.11228 6 Taking into account the economic impacts, the implications for animal health and the possible impact on public health, added to the lack of vaccines or effective treatment, it is extremely important that government entities and policy makers consider adopting a control program for the disease. Once implemented, the control program facilitates the diagnosis of infected animals and allows the adoption of measures that hinder the spread of the virus, minimizing its impact on the cattle production and, eventually, on humans who consume cattle food products (Olaya-Galán et al., 2017). In this context, epidemiological surveys allow the knowledge of epidemiological indicators of infection, guiding control programs and consequently making it possible to control and/or eliminate the virus in cattle populations.
Thus, the objective of the present research was to carry out a survey of the prevalence of enzootic bovine leukosis in a worldwide scenario through a quantitative synthesis with a systematic literature review and meta-analysis.

Methodology
A systematic literature review was carried out with emphasis on the seroprevalence of enzootic bovine leukosis, followed by a meta-analysis of quantitative data available in articles from indexed journals. To prepare the study, the recommendations of the PRISMA methodology -Preferred Reporting Items for Systematic Reviews and Meta-Analyzes To exclude duplicate articles, the tool provided by the bibliographic manager was used. Then, two researchers independently carried out a selection of the studies, initially by analyzing the title and abstract, and later by reading the full papers. After evaluating the texts, other studies were excluded because they did not meet the eligibility criteria. Divergent cases were resolved by consensus.
The survey by Meirelles, Dittrich, Cipriano, & Ollhoff (2009) was subdivided for analysis in three years, in view of the methodology used by the authors. Data extraction was performed individually by two researchers and the information was entered into a previously prepared spreadsheet. Data extracted from the articles were: references (authors and year of publication), type of sampling, total number of animals, number of positive animals, frequency (%) of positive animals, country of study, and diagnostic method.
For the analysis of the quantitative data the 95% confidence interval (95% CI) was

considered. Heterogeneity was assessed by Cochran's Q test and quantified by Higgins and
Thompson's I 2 test. The combined estimates and the 95% CI were calculated based on the random effects model by the inverse of the variance using the DerSimonian-Laird method.
Visual assessment of the funnel plot and the Egger test were also used as alternatives to identify possible biases. All analyzes were performed on the R environment (R CORE TEAM, 2019), RStudio interface (version 1.1.463).

Results
The initial search in the databases and the selection process of the studies are presented in Figure 1. Of the total number of studies surveyed (n = 581) 13 met the eligibility criteria. These papers consisted of cross-sectional surveys (prevalence surveys) with sufficient data for a quantitative synthesis and performance of meta-analysis. The studies included in this stage were carried out in Brazil (n = 2), Bulgaria (n = 1), Canada (n = 1), Chile (n = 1), Colombia (n = 1), United States (n = 1), Iran (n = 1), Iraq (n = 1), Japan (n = 1), Peru (n = 1), Turkey (n = 1), and Venezuela (n = 1). In one of these studies (Meirelles et al., 2009), the prevalence of EBL was analyzed in three different years, being, therefore, considered as three independent studies. Thus, 15 studies were used for the meta-analysis.
The visual analysis of the funnel plot ( Figure 3) showed asymmetric distribution of the 15 studies, showing possible publication biases, however, by applying the Egger test (P = 0.62), the occurrence of such biases was not verified.
In addition, a meta-analysis (Figure 4) was performed considering age as a factor associated with EBL seropositivity, using data from four studies from which it was possible to standardize the age groups of the animals in < 4 years and > 4 years, and the results demonstrated a combined odds ratio (OR) of 2.87 (95% CI = 1.19 -6.93). Research, Society and Development, v. 9, n. 12, e43591211228, 2020 (CC BY 4. Figure 4 -Meta-analysis of four surveys demonstrating the association of age with the sampling, adjustment of rates according to parameters that may influence the prevalence and specific analyzes that provide the determination of reliable epidemiological indicators.