Table 1.

Analysed criteria of the BVDV prevalence studies and summary of these criteria in sub-groups in the meta-analysis and multivariate regression analysis.

Category	Systematic review	Meta-analysis and uni-vs. multivariate regression analysisa
Prevalence	Prevalence data, i.e., percentage and/or the total number of tested and positively tested persistently infected (PI), and/or viraemic (VI) and/or antibody-positive (AB) animals/herds were recorded. If only percentage data were available, then the number of infected animals/herds was extrapolated. N.B. the term BVDV prevalence covers PI, VI and AB prevalence at animal or herd level.	Number of tested BVDV positive animals/herds were included in the uni- and multivariate regression model and included as a weighted measure in the meta-analysis. (1) PI animals/herds: This category covered all animals that have been tested twice antigen (AG) positive within a certain period. In a small number of studies, PI prevalence data were published for one-stage testing or were tested AB-negative and AG-positive. If the timing of testing was such that TI could be excluded (e.g., animals tested after birth before colostrum consumption) or specific diagnostic methods were used such as IHCb testing of tissue samples, animals were further processed as PI instead as VI, although these animals were tested only once. N.B. If at least one PI animal was identified in a herd, the herd was classified as a PI herd. (2) PI/VI animals/herds: If information regarding PI and VI animals was simultaneously available i.e., only some of the VI animals were retested, we only considered the PI prevalence data to avoid double counting of information e.g., with regard to vaccination status of herds or implemented BVDV programmes. In case, that all animals were tested and both test results were negative, we processed these data further as PI prevalence. (3) VI animals/herds: Includes animals tested only once (with the exceptions mentioned under category (1) and (2)) or which were proven to have been TI animals/herds. (4) AB-positive animals/herds: Includes BVDV prevalence data of seropositive infected animals/herds or vaccinated animals.
Country/UN region	Country described the area where cattle were tested. Countries were summarised in respective UN regions (Europe, Australia, West Asia, East Asia, South Asia, North America, South America, Central America, North Africa and Sub-Saharan Africa).	Individual countries assigned to the particular UN region were included.
Period	Date of sampling and the study begin was recorded. If both date of sampling and study begin were not available, this was stated as “not specified”. Additionally, the publication year of the study was recorded to perform Forest Plots.	The sampling periods were summarised in four time periods (before 1992, 1992–2001, 2002–2016, not specified).
Production system	Production systems were categorised into dairy, beef, mixed (i.e., dairy, beef, breeding in mixed populations and others such as stock bulls etc.) and not specified (i.e., cattle).	Production systems were categorised into dairy, beef, mixed and not specified.
Age group	Depending on the age of the sampled animals, the prevalence data should be interpreted with caution as, e.g., calves could be tested AB-positive due to maternal antibodies. If information about the age of animals was not provided, this was stated as “not specified”.	The different ages of the animals were summarised into four age groups: ≤6 months; >6 months; animals in both age groups were classified as “mixed”; and if no age group was available the study was classified in the category “not specified”.
Vaccination	Information on whether the animals/herds tested were vaccinated or not were collected. In case some animals in the herd were vaccinated, the herd was considered vaccinated if more than 10% of animals received immunisation against BVDV. If information regarding vaccination status was not available, the studies were assigned in the category “not specified”.	The vaccination status of the animals and herds was classified into three groups: vaccinated (Yes); not vaccinated (No) and “not specified”.
BVD clinical signs	Clinical signs related to infection such as diarrhoea, respiratory symptoms, reproductive disorders, oral lesions, fever, inflammation of the gastrointestinal mucosae were recorded. If no information regarding clinical signs was provided in the studies, these studies were assigned to the category “not specified”.	The animals and herds were classified into three groups: clinical signs (Yes); no clinical signs (No); and if no information was available the study was classified in the category “not specified”.
BVDV programme	Information regarding control and/or eradication activities were recorded. If no information regarding BVDV programmes was provided, this was stated as “not specified”. BVDV control and/or eradication programmes covered voluntary and/or compulsory BVDV programmes at herd, regional or national level. Control and eradication programmes differ in the degree of disease reduction10, such as control efforts aiming to reduce BVDV prevalence to a relatively low level, while eradication measures aim to provide a continued absence of BVDV10,19. In this study, control programmes did not cover vaccination.	Information about the implemented BVDV programmes were included in the meta-analysis. The information of the BVDV control and/or eradication programmes were distinguished as follows: animals/herds included in BVDV programmes were assigned to the category “Yes”; animals/herds not included in BVDV programmes were assigned to the category “No” and if no information was available, this was covered in the category “not specified”.
Diagnostic method	Types of diagnostic methods used were collected and wherever available the corresponding sensitivity and specificity were recorded.	The different applied diagnostic methods were classified as follows: (1) Direct detection methods comprised IHCb and IFTc; (2) RT-PCRd covered RT-PCR, qPCRe, PCRf, real-time-RT-PCR; (3) AG ELISA covered AG ELISAg or AC ELISAh; (4) Cell culture-based systems covered virus isolation with immunoperoxidase, fluorescent antibody or interference test; (5) Mixed covered combinations of diagnostic methods for screening and confirmation of laboratory results (two-test strategy); (6) AB ELISAi covered AB ELISA, Cell bound immunoassay; (7) NTj covered serum-, virus-, microneutralisation assays; (8) Mixed covered both AB ELISA and NTj; (9) Other covered e.g., IFTc and complement fixation tests.
Sample material	Sample materials such as tissues from living or dead animals except for foetuses and abortion material were recorded. Abortion materials were not incorporated into our analyses because they were considered as unsuitable for diagnostic testing9.	The different sample materials were included in the uni- and multivariate regression analysis and were classified as follows: (1) Blood e.g., whole blood, buffy coat, sera; (2) Blood and milk; (3) Blood and other; (4) Blood and tissue sample; (5) Blood, tissue sample and other; 6) Milk; (7) Tissue sample and other; (8) Tissue sample e.g., ear notch biopsy, brain tissue, spleen, tracheal lymph node, cerebellar tissue; (9) Other e.g., nasal swab, trans-tracheal aspirate, cerebrospinal fluid, rectal swabs.
Level	The studies were categorised into three levels. (1) Farm level: studies covered two to five herds; (2) Regional level: either a certain area/region in a country was sampled or more than five herds (with the exception of feedlot studies, where animals from different areas of a country were investigated); (3) National level: the total cattle population or a representative subpopulation of the total population of one country was tested. Studies not assignable to one of these categories were included in the category “not specified”.	The levels were included in the uni- and multivariate regression analysis.
Genotype	Information regarding to BVDV genotypes 1, 2 and HoBi-like strains was collected.	The genotypes were not included in the meta-analysis and also not incorporated in the uni- and multivariate regression analysis due to insufficient data.

^aUnless otherwise stated, all categories were included in both meta-analysis and multivariate regression analysis.

^bImmunohistochemistry.

^cImmunofluorescence tests.

^dReverse transcription polymerase chain reaction.

^eQuantitative polymerase chain reaction.

^fPolymerase chain reaction.

^gAntigen enzyme-linked immunosorbent assay.

^hA polyclonal antibody-based antigen-capture ELISA.

ⁱAntigen enzyme-linked immunosorbent assay.

^jNeutralisation assays.