Epidemiology of BVD and persistently infected cattle across urban and peri-urban dairy farms in central, northern, and southern parts of Ethiopia Authors: Asamenew Tesfaye Melkamsew, Tesfaye Sisay Tessema, Abebe Garoma, Jan Paeshuyse Published: 2025 ABSTRACT AbstractBackgroundBovine viral diarrhea (BVD) is an economically significant disease affecting both domestic and wild ruminant species. The causative agent is bovine viral diarrhea virus (BVDV) of the genus Pestivirus within the family Flaviviridae. In Ethiopia, information on the disease to initiate prevention and control measures is scarce. Hence, this study aimed to assess the prevalence of BVD and persistently infected cattle and identify the associated risk factors in dairy farms situated in urban and peri-urban areas.MethodologyAn Enzyme-Linked Immunosorbent Assay (ELISA) technique was used to test serum samples collected from 1634 dairy cattle in 180 herds. Univariate analysis was conducted to assess the association between independent and outcome variables, and those with a significant association with the outcome variables were considered for a multivariable binary logistic regression model to identify potential risk factors.ResultsThe study revealed an overall antibody and antigen prevalence of 28.4% (95% confidence interval (CI): 26.2–30.6%) and 1.5% (95% confidence interval: 1.04–2.3%), respectively. The true within- and between-herd- sero-prevalence of BVD was 49.1% and 67.8%, respectively. The study also revealed a 0.94% (95% confidence interval (CI): 0.53–1.77) animal level and 6.4% (95% CI: 0.53–10.9) herd level prevalence of persistently infected animals. The study location, body condition score (BCS), diarrhea, herd size, origin of animals, and calf congenital anomalies were significantly associated with the prevalence of BVD (P < 0.05). Location, BCS, herd size, and origin of the animals were identified as potential risk factors (p < 0.05).ConclusionsOur study revealed that BVD is well established in commercial dairy farms, suggesting the need to design strong prevention and control measures. Future studies should be encouraged to determine the genetic diversity of the viruses in the country.Keywords: BVD, Bovine viral diarrhea virus, Dairy cattle, Ethiopia, Prevalence, PIs, Risk factors Introduction Bovine viral diarrhea (BVD) is a highly contagious, widespread, and economically important disease affecting cattle-producing countries [1]. Recently, BVD has been incorporated into the list of notifiable diseases by the World Organization for Animal Health [2]. It is an enveloped, positive-sense, single-stranded RNA virus called bovine viral diarrhea virus, which belongs to the genus Pestivirus and the family Flaviviridae [3]. The current taxonomic classification based on phylogenetic analysis of conserved amino acid sequences revealed 19 different Pestivirus species. Among species, Pestivirus bovis (BVDV 1) can be further classified into at least seventeen genotypes. There are also Pestivirus tiauri (BVDV 2) and Pestivirus brazilense (HoBi-like virus) [4]. These three species of Pestivirus are attributed to BVD in ruminants. Furthermore, BVDV is classified as cytopathic (cp.) and non-cytopathic (ncp) based on its lytic effect on cell culture [5, 6]. Cattle are highly susceptible to the virus; however, other domestic and wild ruminants, such as sheep, goats, pigs, and buffalo, can also become infected [7]. Although Border Disease Virus (BDV) mainly infects sheep and goats [8], the virus can also infect cattle, which makes differentiation from BVDV during diagnosis is difficult [9]. In cattle, the disease presents in a spectrum of forms, from subclinical infections to those with severe clinical symptoms [10]. The virus exhibits strong tropism towards susceptible animals’ respiratory, reproductive, and gastrointestinal tracts. Furthermore, both transient and persistent BVD infections in cattle can lead to immunosuppression, increasing their susceptibility to secondary infections. The stage of pregnancy plays a critical role in determining disease outcomes for the foetus Infection during early pregnancy may result in embryonic death, abortion, or persistent infection (PI), while infection in late pregnancy can lead to abortion, stillbirth, or severe congenital anomalies [8]. Persistently infected calves are born when a pregnant dam is infected by a non-cytopathic BVDV from day 40 to 125 of pregnancy [11]. Such calves remain immunotolerant, viremic, and shed the virus in all body secretions and excretions throughout their lifetime [1, 12]. The presence of PI animals in dairy herds significantly increases the prevalence of BVD [1]. Persistently infected calves might develop a serious and deadly disease called mucosal disease (MD), which damages the mucosal lining of the gastrointestinal system. Furthermore, many PI calves die prematurely due to respiratory and enteric diseases [13]. Ethiopia’s cattle population is estimated at approximately 70,291,776. Of this total, cross and exotic breeds comprise 2.29% and 0.31%, respectively [14]. The livestock sector contributes approximately 12–16% of the national GDP, 30–35% of the agricultural GDP, 15% of export earnings, and 30% of agricultural employment [15]. Smallholder dairy production systems are the common form of dairy farming in Ethiopia, typically marked by low production and productivity [16]. Several factors are associated with lower production in the sector, such as livestock diseases, low quality and quantity of feed, poor management, lack of infrastructure, and veterinary service provision [17]. In Ethiopia, the annual calf mortality rate ranges from 15.3% to 25% in urban and peri-urban dairy production systems [18]. Diarrhea and pneumonia are identified as the major causes of young stock mortality in dairy farms. Particularly, the premature loss in terms of stillbirth was estimated at approximately 6.2–14.4% [18]. In adult dairy cattle, reproductive diseases, such as repeat breeders, retained fetal membranes, abortion, stillbirth, and infertility, are the major challenges in terms of dairy production [19]. According to Barrett et al. [20], production losses due to high calf mortality and morbidity are positively associated with the seroprevalence of Bovine Viral Diarrhea (BVD). In Ethiopia, there is no vaccination program against BVDV; therefore, detecting BVDV antibody or antigen indicates viral circulation in dairy herds. The objectives of this study were to determine the animal- and herd-level antibody and antigen prevalence in acute(transient) BVDV-infected cattle, prevalence of persistently infected cattle, and assess associated risk factors within urban and peri-urban commercial dairy farming systems in selected study areas of Ethiopia. Methods Description of the study areas The study was carried out in commercial dairy farms found along the dairy value chain of the central (Addis Ababa, Sebeta, Suluta, Sendafa, Bishoftu, and Holeta), southern (Hawassa), and northern (Bahir Dar and Gondar) parts of Ethiopia. Addis Ababa and cities in the Oromia region are considered central Ethiopia (Table 1; Fig. 1). The high potential for dairy production, the presence of a substantial number of exotic and cross-breed dairy cattle, and the diverse agroecological conditions supported the inclusion of these areas in the study. As shown in Table 1the three regions comprise 76.78% of the total cattle population in Of the total estimated 1,828,508 improved and exotic dairy animals, 86.7% were found in these three regions [14]. Table 1The study areas and the geographical locations№RegionsCitiesCattle population+Dairy Cattle*Distance (Km)Altitude (m)LatitudeLongitude1Central EthiopiaAddis Ababa25,506,40991,815-2,3559.0192 °N38.7525 °ESebeta272,3568.9122 °N38.6268 °ESuluta3327959.1853 °N38.7604 °ESendafa38.625149.1496 °N39.0085 °EBishoftu5919208.7346 °N39.0085 °EHoleta2723919,0684 °N38.4935 °E2Southern EthiopiaHawassa11,128,01755,00028217087.047 °N38.4958 °3Northern EthiopiaBahir Dar17,252,804116,000499180011.5896◦N37.3875◦EGondar658213312.6030 °N37.452 L °ENB The distance is taken from the capital city, Addis Ababa. The laboratory is found in Sebeta, which is 24 Km Southwest of Addis Ababa+It includes local, exotic, and cross breeds, *Exotic and/or cross breeds (improved breeds) The study areas and the geographical locations NB The distance is taken from the capital city, Addis Ababa. The laboratory is found in Sebeta, which is 24 Km Southwest of Addis Ababa +It includes local, exotic, and cross breeds, *Exotic and/or cross breeds (improved breeds) Map showing the study areas in Ethiopia from where samples were collected. The panel on the left top represents the northern part (Bahrdar and Gondar), the left bottom represents the southern part (Hawassa), and the right top represents the central part (Addis Ababa, Holeta, Suluta, Sebeta, Bishoftu, and Sendafa) of Ethiopia Study population and herd selection The study animals consisted of cross and exotic-breed dairy cattle of both sexes, aged > six months, reared in urban and peri-urban commercial and smallholder dairy farms. Body condition is categorized as poor, moderate, good, fat, or grossly fat [21]. Depending on herd size, dairy farms were categorized as small (5–10 cattle), medium (11–25 cattle), or large (> 25 cattle. Study design and sample size determination A cross-sectional study was conducted between August 2019 and March 2023 across urban and peri-urban dairy farms, to determine the prevalence of BVD and PIs, as well as to determine associated risk factors. study areas. Previously registered dairy farms under the Ethiopian Control of Bovine Tuberculosis (Ethicobots) project were used as the sampling frame, and the dairy farms were randomly selected in each region as defined in Table 1. A two-stage cluster sampling technique was used to determine the number of herds and individual animals needed for the study [22]. The primary unit was the herd, selected using a random number generator from Epitools [23], and the secondary unit was individual animals, chosen through simple random sampling (lottery method) for farms with animal identification and systematic random sampling for those without identification. Previous individual-level prevalence of 28.4% [24], a degree of precision of 0.1, the number of animals sampled per herd [(Small Scale (SSC) = 5, Medium Scale (MSC) = 13, and Large Scale (LSC) = 25)], and an intra-cluster correlation coefficient variation (rho) of 0.2 were taken into account. Accordingly, a total of 796 cattle from 119 herds were sampled in Central Ethiopia, including Addis Ababa. Additionally, 435 cattle from 33 herds were sampled in Northern Ethiopia, and 403 cattle from 28 herds in Southern Ethiopia. In total, the study included 1,634 dairy cattle from 180 herds. Sample collection Blood samples were collected by puncturing the jugular vein using vacutainer needles (18 gauge) and ethylenediaminetetraacetic acid (EDTA)-free blood collection tubes. From each selected animal, 5–6 mL of blood was drawn, and the blood was kept for 15 min at room temperature to allow clotting. The clot was removed by centrifuging at 2500 rotations per minute for 15 min at 4 °C. Serum samples were harvested into labeled sterile cryovial tubes and transported to the Animal Health Institute (AHI) using a cool box with an ice pack for laboratory analysis. Upon arrival, the samples were stored at −20 °C until laboratory analysis was conducted. Laboratory analysis Antibody detection To detect anti-BVDV antibodies in the serum samples from acute(transient) BVDV-infected cattle, a BVDV/MD/BDV p80 antibody test kit (IDEXX Laboratories, Inc., Westbrook, Maine 04092, USA) was used. The serological kit has a sensitivity and specificity of 60% and 97%, respectively [25]. The serological analysis was carried out following the manufacturer’s instructions. The results were obtained by computing the sample absorbance with the mean absorbance of the negative control, as follows. The results were interpreted as positive, doubtful, or negative whenever the S/N% is ≤ 40, 40 ≤ S/N < 50, or ≥ 50%, respectively. The test was valid if the mean OD of the negative control (NC) was ≥ 0.8, and the ratio of the positive control (PC) to the mean of NC was < 0.2. Antigen detection To detect bovine viral antigen (BVDV), Bovine Viral Diarrhea Virus (BVDV) Erns Antigen Test Kit/serum plus (IDEXX Laboratories, Inc., Westbrook, Maine04092, USA) was used according to the manufacturer’s instructions. Animals that became negative for antibody and positive during antigen detection were considered to bePI animals [24]. The presence or absence of BVDV antigen in the sample was determined using the corrected OD value (Sample-Negative Control) for each sample. Finally, the result was interpreted as negative and positive if S‒N ≤ 0.3 and S‒N > 0.3, respectively. The test was valid if the mean OD of the negative control was ≤ 0.25, and the difference between the mean OD of the positive control and the negative control was ≥ 0.15. Definitions Acute (transient) BVDV infection is encountered more frequently in young animals. It may be clinically inapparent or associated with fever, diarrhoea, respiratory disease, and sometimes sudden death, followed by clearance of the virus and the development of immunity [9]. Persistent infections(PIs) are viremic cattle throughout their life, while seronegative to the BVDV species to which they are exposed [9]. Animal-level antibody or antigen prevalence is the proportion of individuals in a herd or population that test positive for antibodies against BVDV or are BVDV antigen positive at a specific time or during a specified period. Herd-level antibody or antigen prevalence is a proportion of herds in which at least one animal is positive for antibody against BVDV or is BVDV antigen positive at a specific time or during a specified period. Statistical analysis The data generated were entered, coded, and stored in Microsoft Excel 2007 for Windows 2010. The data were imported into SPSS version 20 (SPSS Inc., Chicago, IL, USA) for analysis. A descriptive analysis was conducted on the serological survey. The true and apparent prevalence, along with the within and between-herd-level prevalence, were calculated according to Brown and Cai [26] and Rogan and Gladen [27]. Pearson’s Chi-square test was used to determine significant differences in dichotomous predictor variables between outcome groups. Differences among the groups for each factor were considered significant at p < 0.05. Before the regression analysis, the potential predictors were analyzed individually using univariate analysis for association with the outcome at p < 0.25 to determine predictor variables to include initially in the model. Multi-collinearity tests using Variance inflation factors (VIF) less than 10 and Tolerance value greater than 0.1 [28]. Multivariable binary logistic regression analyses used to estimate the odds ratio (OR), which quantifiesthe strength of the association between the outcome (dependent variable) and explanatory variables (predictors or independent variables). Hosmer and Lemeshow goodness-of-fit was used to evaluate how well the regression model matches the observed data, and p > 0.05 indicates no evidence of poor fit (good model fit), and p ≤ 0.05 suggests the model does not fit the data well [29]. Finally, the adjusted odds ratio with 95% CI at p < 0.05 was considered to include predictor variables as potential risk factors for BVDV seroconversion. Results Overall antibody prevalence Overall, out of 1634 animals, 464 were serologically positive, providing an animal-level prevalence of 28.4% (95% confidence interval (ci): 26.2–30.6). An apparent prevalence of 40.2% and 65.3% true prevalence were recorded in the northern part (Table 2), the highest prevalence compared to the south and central parts of Ethiopia. Table 2The apparent and true antibody prevalence of BVDLocationNo. TestedNo. PositiveApparent prevalenceTrue prevalenceEstimate (%)95% CIEstimate (%)95%CICentral Ethiopia79613116.514.1–19.223.719.4–28.4Northern Ethiopia43517540.235.7–49.965. 357.4–74.0Southern Ethiopia40315839.234.6–44.164.055.4–72.0Total163446428.426.3–30.644.640.8–48.5 The apparent and true antibody prevalence of BVD It can be seen from the data in Table 3 location (ꭕ2 = 108.9; p = 0.00), body condition score (ꭕ2 = 30.2; p = 0.00), profuse diarrhea (ꭕ2 = 5.16; p = 0.023), herd size (ꭕ2 = 8.12; p = 0.017), origin of the animal (ꭕ2 = 8.423; p = 0.004), and existence of calf abnormality (ꭕ2 = 4.78;p = 0.029) were significantly associated with the antibody prevalence of BVD. Table 3Prevalence of BVD antibody? And its association with different predictor variablesVariables№ Sampled№ Negative№ PositivePrevalence (%)95%CIꭕ2-valueP valueRegionsCentral Ethiopia79669513116.514.0–19.1.0.1108.90.000Northern Ethiopia43526017540.235.6–44.8Southern Ethiopia40324515839.234.4–44.0Total1634117046428.426.3–30.6SexFemale1549110944028.426.2–30.1Male86612428.218.7–37.70.010.973Total1634117046428.426.2–30.1Age (year)0.5–1.5347250972823.3–32.71–23012297223.919.1–28.74.120.127> 298669129529.927.0–32.8.0.8Total1634117046428.426.2–30.1BCSPoor57446610818.8.315.8–22.244.60.000Good76051924131.728.5–35.1Fat30018511538.333.0–44.0Total1634117046428.426.2–30.13.180.204ParityNulliparous44934410523.419.7–27.58.60.013Primiparous2471777028.323.1–34.3Multiparous62142519631.628.0–35.3.0.3Total131794637128.225.8–30.1Profuse diarrheaNo1550111943127.825.6–30.05.160.023Yes84513339.328.9–49.7Total1634117046428.426.2–30.1Respiratory problemsNo1571113044128.125.9–30.32.120.145Yes63402336.524.6–48.4Total1634117045428.426.2–30.1HerdSSC56411526.815.2–38.48.120.017SizeMSC79433645.634.6–56.6LSC45212453.338.7–67.9Total1801057541.726.2–30.1OriginHome95653031.622.3–40.98.4230.004Purchased85404552.942.3–63.5Total1801057541.726.2–30.1BreedingNatural2617934.616.3–52.92.780.249AI112684439.330.3–48.3Both42202252.437.3–67.5Total1801057541.726.2–30.1Bull sharingNo53302343.431.- 56.70.920.868Yes127755240.932.−49.6Total1805312770.663.5–76.7Calf abnormalityNo117384732627.825.2–30.44.780.029Yes36201644.428.2–60.6Total120986734228.328.8–30.8AbortionNo1135111531627.825.2–30.41.8220.177yes74732635.124.2–46.0Total1209118834228.328.8–30.8NB. SSC Small Scale, MSC Medium Scale, LSC Large Scale, BCS Body Condition Score, AI Artificial Insemination Prevalence of BVD antibody? And its association with different predictor variables NB. SSC Small Scale, MSC Medium Scale, LSC Large Scale, BCS Body Condition Score, AI Artificial Insemination Based on the univariate analysis, the variables age (p = 0.283), sex (p = 0.677), and sharing bull (p = 0.656) were found to be non-significant. A variable, such as profuse diarrhea (p = 0.021) and ocular or nasal discharge (p = 0.173), was removed from the analysis because they are not risk factors but effects of the disease. The results of multivariable binary logistic regression showed that the study location, herd size, parity, BCS, and origin of the animal were identified as potential risk factors (p < 0.05). Similarly, animals having poor and good body condition had about 2.68 and 1.55 more likely to be positive than fat animals, respectively. Moreover, those dairy farms introducing new animals from the market had 2.44 times greater odds of BVD than closed breeding farms. However, animals kept in SSC dairy farms had 0.38 times less odds of BVD than animals managed under LSC dairy farms. Similarly, nulliparous animals were 0.71 times less likely to become BVD positive than multiparous animals (Table 4). Table 4Potential risk factors based on multivariable binary logistic regressionVariables№ Sampled№ Negative№ PositiveOdd ratioP value95%CILocationCentral Ethiopia7966951310.310.0000.23–0.40Northern Ethiopia4352601751.00.790.79–1.34Southern Ethiopia40324515810.980.62–1.64BSCPoor5744661082.680.0001.96–3.67Good7605192411.550.0001.55–2.6Fat3001851151ParityNulliparous4493441050.710.0030.5–0.87Primiparous247177700.860.350.6–1.19Multiparous6214251961HerdSSC5641150.320.0070.14–0.74MSC7943360.730.410.35–1.53LSC4521241OriginPurchased8540452.440.0041.33–4.47Home95653010.760.58–2.11NB. SSC Small Scale, MSC Medium Scale, LSC Large Scale, BCS Body Condition Score, AI Artificial Insemination Potential risk factors based on multivariable binary logistic regression NB. SSC Small Scale, MSC Medium Scale, LSC Large Scale, BCS Body Condition Score, AI Artificial Insemination Overall antigen prevalence Of the 1634 animals tested, Bovine viral diarrhea virus antigens were detected in 25 samples, providing an animal-level prevalence of 1.5% (95% confidence interval (ci):1.04–2.3). All antigen-positive animals originated from central Ethiopia. The virus was detected in animals aged between 0.75 and 17 years, with higher prevalence (1.1%) observed in animals aged three years and older The overall antigen prevalence (1.41%) was higher in female animals (n = 23). Among these, nulliparous or primiparous, representing 0.5% of the sample. The antigen prevalence was also higher on farms with homebred cattle (1.41%). Additionally, farms utilizing natural and artificial breeding methods showed a prevalence rate of 0.67%. The prevalence of persistently infected (PI) animals with BVDV was 0.67% (n = 11; 95% confidence interval [CI]: 0.38–1.2%). A higher prevalence (0.43%) was observed in animals aged two years or younger. Among female and nulliparous animals, PI prevalence was 0.49% (8/1634) and 0.31% (5/1634), respectively. Out of 180 herds, five tested positive for PI animals, resulting in a herd-level prevalence of 2.8%. Similarly, farms utilizing both natural and artificial breeding methods recorded a PI prevalence of 2.8% (5/180). Herd level prevalence A total of 41.7% (95% confidence interval (CI): 26.2–30.1) of these herds were found to be positive for antibodies against BVDV, of which 53.3% were from animals kept in LSC dairy farms (Table 5). Table 5The apparent and true antibody prevalence of BVD at the animal and herd levelPrevalenceNo. TestedNo. PositiveApparent prevalenceTrue prevalenceEstimate (%)95%CIEstimate (%)95%CIAnimal level163446428.426.3–30.64440.8.−48.5Within herd13424163128.6–33.549.144.9–53.6Between herd1807541.734.7–49.067.855.6–80.7 The apparent and true antibody prevalence of BVD at the animal and herd level Discussion Although the availability and use of various BVDV vaccines, BVD outbreaks have been reported globally [30]. BVD is an economically significant disease in cattle-producing countries; however, information on BVD prevalence in Ethiopia remains insufficient to initiate prevention and control programs. The present study revealed an overall apparent prevalence of 28.4% for BVDV antibodies and 1.5% for BVDV antigens at the animal level. An antibody prevalence of 41.7% was recorded at the herd level. The prevalence of persistently infected (PI) cattle was estimated at 0.67% at the animal level and 2.8% at the herd level, suggesting that the majority of the animals were transiently infected with BVDV. The findings clearly indicate active circulation of the virus in the herds, given the absence of a BVDV vaccination program in the country. However, due to limitations of ELISA techniques, such a potential for cross-reaction with other Pestivirus species; further confirmation using molecular and viral neutralization techniques is required. Almost all dairy cattle sampled in this study had no contact with small ruminants, which are known reservoirs of Border Disease Virus (BDV). In the few instances where small ruminants were present on dairy farms, samples were collected and tested for both BVDV and BDV to minimize the risk of cross-reactivity in diagnostic results. Fortunately, all samples from small ruminants tested negative for both viruses. This supports our conclusion that BVDV is circulating with a high prevalence in the country. Furthermore, molecular investigations conducted in the same areas and dairy farms confirmed that BVDV genotypes 2a and 1b are predominant in the study regions [31].The animal-level antibody prevalence in this study was within the range of 10.7–32.9% reported by earlier studies in different parts of Ethiopia [32, 33]. This study supports previous findings of higher prevalence in other African countries 19.8%–79.1% in Kenya [34–36], 37.71% in Morocco [37], and 66.39% in Nigeria [38]. Our finding is contrary to those of Tesfaye et al. [39]and Tadesse et al. [40], who reported 80.82% and 51.7% prevalence in Borena and Somali pastoral areas, and southwest Ethiopia, respectively. The overall prevalence was higher than that reported previously from the same study location [32]. The variation in antibody prevalence might be due to differences in biosecurity measures, sensitivity and specificity of the test, herd size, age at sampling, farm management practices, differences in sample size, presence of PI animals, and seasonal variation [41]. The results of this study revealed that the antibody prevalence was significantly associated with study location. This is consistent with a previous report that showed the relationship between BVD and geographical locations [34, 42]. Moreover, it also confirms the same report elsewhere [42, 43]. The current study found that the antibody prevalence varies significantly among age groups. A comparison of these findings with those of other studies confirms higher seroconversion in adult cattle [41, 44]. Moreover, this study revealed that the higher herd-level antibody prevalence was significantly associated with herd size. This result corroborates that of Aragaw et al. [44], who reported a 50% prevalence of BVD in Ethiopia. Consistent with this result, Sarrazin et al. [45] reported a higher antibody prevalence of BVD in large-scale dairy farms. However, these findings were lower than the 68.3% [24] and 95.6% [40] antibody prevalence reported in previous studies from the same country. It is also lower than the 77.8% prevalence report from Eastern China [46]. A lower prevalence (22%) of the current finding was also reported from Colombia [47]. The variation in disease prevalence might be associated with the repeated exposure of adult animals to the virus; as a result, antibodies persist throughout the animal’s life [48]. Furthermore, the management practice, high cattle density, and the existence of PIs increase the herd-level antibody prevalence of BVD [45]. The current investigation revealed that the antibody prevalence of BVD significantly varies between the origin of the animals. These results reflect those of Graham et al. [49], who reported that introducing purchased animals to the herd significantly contributes to increased seropositivity for BVD. This finding supports the work of other studies showing trade was recognized as an important means of spreading infection between herds through the movement of PI and/or transiently infected animals [50]. The risk of introducing disease to herds through purchasing animals is well recognized [51, 52], hence increasing the chance of BVDV-free herds being infected with the virus. This study found that the overall antigen prevalence was 1.5% in dairy farms. This finding supports the work of other studies, such as a 1.8% prevalence from Brazil [53] and a 2.2% prevalence from Korea [54]. The antigen prevalence in this study was higher compared to reports from Botswana [55] and Croatia [56]. Our finding was contrary to Yitagesu et al. [57], who reported 0% prevalence in Ethiopia. However, the prevalence was lower than the 3.2% reported from Iran [58] and 10.7% from Sudan [59]. The prevalence of BVD antigens varies significantly with the body condition of dairy animals. With the present results, the previous study also demonstrated the same result [60]. Thomas [61] confirmed in their experimental study that persistent infection with BVDV is positively correlated with poor body condition score. Therefore, poor body condition scoring could be an indicator for PI animals [60, 62]. The poor body condition observed in PIs is primarily due to the continuous viral replication, which damages the gastrointestinal tract and blocks nutrient absorption and general metabolism. Furthermore, the chronic diarrhea commonly seen in PI animals further contributes to weight loss and poor body condition. A possible explanation for the variation in antigen prevalence may be associated with sample type, sensitivity, and specificity of diagnostic techniques, and the differences in farm management. Interestingly, for the first time, we reported a 0.67% prevalence of persistently infected animals from the studied areas. The finding in this study is below the global PIs prevalence (0.8 to 1.6%) [30]. This result was contrary to Yitagesu et al. [57], who found 0% prevalence in Ethiopia. Comparison of the findings with those of other studies confirms that the prevalence was higher than 0.13% in Brazil [63], and 0.32% in Belgium [64]. The prevalence in this investigation was lower compared to 4.1% in Uruguay [65], This discrepancy could be attributed to the variation in sample size and sampling strategies, herd size, sensitivity and specificity of diagnostic techniques, disease prevention and control strategies, and production system. In this study, the herd-level prevalence for persistently infected animals accounted for 6.4%. This result seems supported by a 7% herd-level prevalence reported by Quintero et al. [47]. The prevalence was lower than 37.5% in Brazil [66], and 22% in Colombia [47]. The same justification indicated for animal-level prevalence could be used for herd-level prevalence. Conclusions The present study aims to determine the prevalence of BVDV antibodies and antigens in randomly selected dairy farms across central, northern, and southern Ethiopia. One of the most significant findings to emerge from this study is the prevalence report for persistently infected animals. Several potential risk factors, such as location, age, BSC, herd size, and origin or source of the animal, are identified as the main risk factors for a higher prevalence of BVD. Taken together, these results suggest that BVD is prevalent in commercial dairy farms. More broadly, to address the limitations of serological analysis (ELISA), broader research employing molecular techniques is needed to determine the genetic diversity of BVDV circulating in the country. Acknowledgements This research work was made possible by the support of the VLIR-UOS Global Minds program and the members of the Host-pathogen interactions and the staff at the National Animal Health Institute. The authors appreciate the support staff at the Animal Health Institute in terms of sample collection and laboratory analysis. We also thank Professor Teshal Sori and Dr. Hassen Chaka for reviewing this manuscript. Authors’ contributions A.T. was involved in the conception of the study and design, data analysis and writing—original manuscript. Moreover, in sampling and laboratory analysis. T.S. participated in the conception of the study and design, review and editing of the manuscript. A.G. conducted sampling and laboratory analysis. J. P. was involved in the conception of the study and design, review and editing, funding acquisition, and supervision. Funding This project was fully supported by the VLIR-UOS through the Global Minds program at KU Leuven in Belgium. Data availability All data generated and analyzed during this study are included as tables in the manuscript. However, the raw data are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate Ethical clearance was obtained from the Animal Health Research Scientific Ethical Committee of Animal Health Institute (AHI), Ethiopia. All methods were performed in accordance with the international animal care and use guidelines adopted by the Ethical Committee in AHI. Before conducting the research, dairy farm owners were informed of the objectives of the study, and verbal consent was obtained from the dairy farm owner or manager. Informed consent was received from all the willing participating farmers and farm managers. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Asamenew Tesfaye Melkamsew, Email: asamenewtesfaye.melkamsew@kuleuven.be, Email: asefiker@yahoo.com. Jan Paeshuyse, Email: jan.paeshuyse@kuleuven.be. References Associated Data Data Availability Statement All data generated and analyzed during this study are included as tables in the manuscript. However, the raw data are available from the corresponding author upon reasonable request.