As early as 1954, the theory that leukosis in cattle could be associated with a transmissible virus was presented.1 By 1973, further evidence had associated a specific virus with bovine lymphosarcoma.2–4 Bovine leukemia virus (BLV), an oncogenic retrovirus of the genus Deltaretrovirus, infects bovine lymphocytes and is the etiologic agent of bovine leukosis. A review of natural transmission of BLV in dairy and beef cattle reported an infective dose of viable BLV-infected lymphocytes can be transmitted from BLV-infected cattle to BLV-free cattle by the transmission of lymphocytes through blood, secretions, or excretions.5 Herd-level probability of the presence of BLV-infected cattle and within-herd seroprevalence in positive herds has not been well characterized for US beef cattle. The 1997 US Department of Agriculture National Animal Health Monitoring System (USDA NAHMS) survey reported that 38.7% of beef operations had at least 1 seropositive animal and 10.3% of sampled beef cows were seropositive, but the report cautioned that the number of herds sampled was insufficient to provide a national estimate of BLV infection and the current seroprevalence is likely different than that reported 25 years ago.6 While some limited-scope studies did not provide evidence that the detection of BLV antibodies in blood was associated with beef cow longevity over a 2-year monitoring period7 nor that natural breeding with BLV-positive bulls was associated with BLV transmission,8 much remains to be learned about the potential effect of BLV infection on the health and productivity of US beef cattle herds.
The status of cows relative to infection with and infectivity of BLV can be assessed in several ways. ELISA of serum samples detects antibodies to BLV, interpreted as prior infection with the virus that elicited an antibody response. Because BLV infection is currently considered to be lifelong, the presence of BLV antibodies is interpreted as a current BLV infection. BLV will integrate within the genome of infected cells as a provirus, but provirus is not detected with quantitative polymerase chain reaction (qPCR) tests in all blood samples from ELISA-positive cattle.9 Differentiating ELISA-positive but qPCR-negative cattle from qPCR-positive cattle may be important to understand the epidemiology of BLV if ELISA-positive but qPCR-negative cattle represent cattle with latent BLV infections are not actively infectious. Some researchers have identified that among cattle with provirus detectable by qPCR, the number of proviral copies per infected cell ranges greatly with most cattle expressing low copy counts, which leads to hypotheses that the small percentage of cattle expressing high proviral load (PVL) may be disproportionately important in the transmission and/or progression of BLV.9–12 Although many cows infected with BLV do not manifest clinical or hematologic signs of disease, some BLV-infected cows develop persistent lymphocytosis and have high absolute lymphocyte counts13,14, and total WBC concentrations may be correlated with BLV PVL.15
The objective of our cross-sectional study was to utilize a convenience sample of Eastern Kansas beef cow-calf herds to test the association between BLV ELISA-status, qPCR-status, and PVL-status with the probability of being classified as pregnant and the probability of being classified as becoming pregnant in the first 21 days of the breeding season.
Materials and Methods
Study population
Forty-four beef cow-calf herds located across 13 counties in Eastern Kansas were enrolled in the study. For this study, a herd was defined as a group of cows managed as a unit through the grazing season. To model the effects of study variables, 1 herd with a clearly disparate pregnancy percentage was excluded (Figure 1). The excluded herd had 25 3-year-old cows of which 96% of the cows were BCS 4, 48% were BLV ELISA-positive, and 32% (8/25) were pregnant. Therefore, the study evaluated 43 herds with a total of 2,820 cows with herd sizes ranging from 21 to 202 cows.
Count of original 44 herds classified by pregnancy percentage decile.
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
Count of original 44 herds classified by pregnancy percentage decile.
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
Count of original 44 herds classified by pregnancy percentage decile.
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
The 43 herds were nested in 19 ranches (ie, some ranches had more than 1 herd, n = 1 to 6 herds per ranch). To maintain the anonymity of the producers and their cattle, each ranch was assigned a unique number, and each herd within a producer’s ranch was assigned a herd number. All study procedures involving cattle were conducted in accordance with a protocol approved by the Kansas State University Institutional Animal Care and Use Committee (IACUC #4626). The survey instrument submitted to the Kansas State University Institutional Review Board was deemed exempt (IRB #10802).
Data collection
Beginning on September 27, 2021, and ending on November 5, 2021, pregnancy-status data were collected by veterinarians from the College of Veterinary Medicine, Kansas State University (S.H., J.A., and B.M.W.) and recorded in a Microsoft Excel spreadsheet. Fetal age estimation was determined by uterine palpation or ultrasonic examination per rectum and categorized into 21-day intervals.16–19 Body Condition Score (BCS) are categories used to describe the relative fatness or body fat reserves of a beef cow using a range of 1 to 9, with a score of 1 representing a very thin cow and 9 representing an extremely obese animal.20 Cow age was provided by herd records and BCS was collected by S.H. and B.M.W. at the time of pregnancy diagnosis.
Whole blood was collected into a 6 mL EDTA (purple top) vacutainer and a 6 mL non-additive (red top) vacutainer via caudal tail vein. Blood was kept on wet ice and transported to the KSU College of Veterinary Medicine in Manhattan, KS, within 8 to 12 hours after collection. The blood samples were then immediately logged and the non-additive vacutainers were centrifuged at 3,000 RPM for 15 minutes to isolate serum.
Serum ELISA determination of BLV-status
Serum was submitted to the Kansas State Veterinary Diagnostic Laboratory (KSVDL) for a BLV ELISA antibody test (gp-51; IDEXX).
Serum qPCR determination of BLV-status
The EDTA vacutainers were refrigerated and next-day shipped with icepacks for BLV qPCR testing (SS1 qPCR test; CentralStar Laboratories). Because of budgetary constraints, qPCR testing of all cows was not possible. We elected to exclude cows that were identified as ELISA-negative for qPCR testing because we assumed that ELISA-negative cows were unlikely to be qPCR-positive. For samples that were collected on Monday through Thursday, qPCR testing was only completed on cows that were found to be ELISA-positive. Our concern for DNA degradation due to a delay in obtaining ELISA results for Friday samples that were submitted to the KSVDL on Saturday mornings led us to submit all samples collected on Fridays for qPCR testing upon arrival to CentralStar Laboratories, providing complete ELISA and qPCR data on 14 herds (724 cows).
Serum PVL determination of BLV-status
The SS1 qPCR measures the ratio of proviral copies per host cell DNA.21 This ratio is an estimate of the percentage of lymphocytes infected with the virus. The mean and median PVL for all qPCR-positive animals in this study was 0.73 and 0.50 proviral copies per host DNA, respectively. Currently, there is no recognized and consistent method for reporting PVL, and various cut-offs for culling or classification as “high-PVL” have been published.7,11 For this study, we defined high-PVL as ≥0.9 proviral copies per host DNA. We selected 0.9 because this cut-off results in approximately 10% (276/2,820) of the study population being classified as high-PVL if all untested ELISA-negative cows are assumed to have 0 PVL.
Data cleaning
Data were entered into a spreadsheet (Excel; Microsoft Corp.) and subsequently analyzed using R Studio (R-Core, 2020). Data were cleaned, described, and re-categorized using the dplyr package22 and graphs developed using ggplot2.23 Cow ages were collapsed into 4 age categories: primiparous cows (heifers), cows that were 2 or 3 years of age, cows that were 4 through 9 years of age (mature), or cows that were 10 years of age or older (old). These age classifications were based on the authors’ impression that the probability of becoming pregnant has clinically meaningful variability between these classifications.
Statistical analysis
A total of 6 multivariable logistic regression models were used to evaluate associations of the 3 methods of classifying BLV-status (ELISA-status, qPCR-status, and PVL-status) with the 2 classifications of pregnancy-status (overall probability of pregnancy and probability of becoming pregnant in the first 21 days of the breeding season) while controlling for other covariates.
For each multivariable logistic regression model, pregnancy status was modeled as a binary outcome (1 = pregnant, 0 = not pregnant; or 1 = pregnant in the first 21 days, 0 = not-pregnant within the first 21 days), and the herd was nested within ranch as a random effect. Each model began with a fixed effect of BLV-status based on 1 of 3 classification methods with BLV ELISA-status classified as a binomial positive or negative, and BLV qPCR PVL results classified in 2 ways; the first, as a binomial positive or negative, the second, as binomial high-PVL (≥0.9) or low-PVL (<0.9). Other potential covariates included in the original models were the age category and BCS category. Interactions of age category by BLV-status and BCS category by BLV-status were also included in the original models. BLV-status was forced into the model regardless of P-value, and the cut-off for potential covariates to be included in the final models as fixed effects was a P < .05. Explanatory variables that failed to meet the cut-off threshold were removed from the model using a backward elimination methodology.
Results
Description of study herds
The herd-level probability of pregnancy for the 43 herds included in the analysis ranged from 0.77 to 1.0. In addition, the probability for cows in each of the 43 study herds to be classified as becoming pregnant in the first 21 days of the breeding season ranged from 0.04 to 0.80 (Figure 2). Herds having 40 or fewer cows accounted for 37.2% of the data set and 24.1% of the cows in the study were classified as heifers. Cows in this data set were all classified as BCS 4, 5, 6, or 7 resulting in no cows being classified as BCS 1, 2, 3, 8, or 9, and 62.4% of the cows were classified as BCS 5. Of the 43 study herds, 95.3% (41/43) had a least 1 ELISA-positive cow.
Distributions of study herds’ (n = 43) pregnancy percentage in the first 21 days of breeding (2A), study herds’ sizes (2B), study cows’ (2,820) age categories (2C), and study cows’ body condition score (BCS) categories (2D).
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
Distributions of study herds’ (n = 43) pregnancy percentage in the first 21 days of breeding (2A), study herds’ sizes (2B), study cows’ (2,820) age categories (2C), and study cows’ body condition score (BCS) categories (2D).
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
Distributions of study herds’ (n = 43) pregnancy percentage in the first 21 days of breeding (2A), study herds’ sizes (2B), study cows’ (2,820) age categories (2C), and study cows’ body condition score (BCS) categories (2D).
Citation: American Journal of Veterinary Research 84, 5; 10.2460/ajvr.22.09.0157
Three methods of classifying BLV-status of beef cows
Raw data from the 2,820 cows in the 43 study herds reveals that 55% (1,552/2,820) were classified as BLV-positive by ELISA. Blood samples from nearly all (1,551/1,552) ELISA-positive cows in the study were evaluated with a qPCR test (SS1 qPCR test; CentralStar Laboratories). The blood sample from 1 ELISA-positive cow was not available for qPCR analysis. Fifty-four percent of the ELISA-positive cows (832/1,551) were classified as qPCR-positive. Of the 164 ELISA-negative cows that were evaluated by qPCR, 97.6% (160/164) were classified as qPCR-negative; therefore, for the analysis, all ELISA-negative cows that were not qPCR-tested were classified as qPCR-negative, and PVL <0.9. We made the decision to tentatively accept the assumption that ELISA-negative cows were likely to be PCR-negative because, constrained by budget limitations, we prioritized an investigation of the association between ELISA-positive cows and PCR-status. More work needs to be done to test the relationship between ELISA-negative cows and PCR-status.
Classifying all ELISA-negative cows that were not qPCR-tested as qPCR-negative resulted in 29.7% (836/2,819) of the study cows being classified as BLV qPCR-positive, and 9.8% (276/2,819) were classified as having a PVL of 0.9 or greater. Among cows classified as BLV-positive by ELISA, 53.6% (832/1,151) were classified as BLV-positive by qPCR and 24.0% (276/1,151) had a PVL of 0.9 or greater.
Probability of becoming pregnant during the breeding season based on BLV-status
The final multivariable logistic regression analysis used to test the association of BLV ELISA-status with the probability of becoming pregnant during the breeding season after backward elimination included BCS-category. The model-adjusted probability of becoming pregnant during the breeding season was statistically significantly associated with BLV ELISA-status (mean [SE]: ELISA-positive = 0.92 [0.05]; ELISA-negative = 0.88 [0.07]; P = .04; Table 1).
Model covariates and model-adjusted pregnancy probability associated with 3 methods of classifying BLV-status of beef cows.
| Fertility outcome = overall probability of pregnancy | ||||
|---|---|---|---|---|
| Method used for determining BLV-status | Covariates included in final model | Pregnancy probability for BLV-positive (PLV-high) cows (SE) | Pregnancy probability for BLV-negative (PVL-low) cows (SE) | P-value |
| ELISA (binomial: positive/negative) | BCS category | 0.92 (0.05) | 0.88 (0.07) | .04 |
| qPCR (binomial: (positive/negative) | Age-category and BCS category | 0.92 (0.05) | 0.90 (0.06) | .13 |
| PVL (binomial: high-PVL/low-PVL)* | Age-category and BCS category | 0.91 (0.05) | 0.90 (0.06) | .86 |
BLV = bovine leukemia virus. BCS = Body Condition Score. qPCR = quantitative polymerase chain reaction. PVL = proviral load.
high-PVL ≥ 0.9 proviral copies/host DNA; low-PVL < 0.9 proviral copies/host DNA.
The final multivariable logistic regression analysis revealed that the model-adjusted probability of becoming pregnant during the breeding season was not significantly associated with BLV qPCR-status (mean [SE]): qPCR-positive = 0.92 [0.05]; qPCR-negative = 0.90 [0.06]; P = .13) when using the assumption that untested ELISA-negative cows were also qPCR-negative (Table 1).
When testing the association of BLV qPCR-status with the probability of becoming pregnant during the breeding season we recognize that the percentage of cows evaluated for PVL (1,551 ELISA-positive cows and 164 ELISA-negative cows) that were classified as being “high” using the cut-off of ≥0.9 was 16.1% (276/1,715), but this comparison over-emphasizes ELISA-positive cows because the PVL of many ELISA-negative cows was not tested. If ELISA-negative cows are assumed to have a PVL equal to 0, the percentage of cows classified as being high using the cut-off of ≥0.9 was 9.8% (276/2,819). The final multivariable logistic regression analysis used to test the association of BLV PVL-status with the probability of becoming pregnant during the breeding season revealed that the model-adjusted probability of becoming pregnant during the breeding season was not significantly associated with BLV PVL-status (mean [SE]): PVL-high = 0.91 [0.05]; PVL-low = 0.90 [0.06]; P = .86) when using the assumption that untested ELISA-negative cows were PVL-low (Table 1).
Probability of becoming pregnant in the first 21 days of the breeding season based on BLV-status
The multivariable logistic regression analyses were used to test the associations for all 3 methods for determining BLV-status with the probability of being classified as becoming pregnant in the first 21 days of the breeding season after backward elimination included age-category but not BCS-category or either of the interactions tested. None of the methods of classifying BLV-status were associated with the probability of becoming pregnant in the first 21 days of the breeding season (Table 2).
Model covariates and model-adjusted probability of becoming pregnant in the first 21 days of the breeding season associated with 3 methods of classifying BLV-status of beef cows.
| Fertility outcome = probability of becoming pregnant in first 21 days | ||||
|---|---|---|---|---|
| Method used for determining BLV-status | Covariates included in final model | Pregnancy probability for BLV-positive (PLV-high) cows (SE) | Pregnancy probability for BLV-negative (PVL-low) cows (SE) | P-value |
| ELISA (binomial: positive/negative) | Age-category | 0.19 (0.07) | 0.19 (0.07) | .99 |
| qPCR (binomial: (positive/negative) | Age-category | 0.18 (0.07) | 0.19 (0.07) | .65 |
| PVL (binomial: high-PVL/low-PVL)* | Age-category | 0.19 (0.07) | 0.17 (0.07) | .49 |
high-PVL ≥ 0.9 proviral copies/host DNA; low-PVL < 0.9 proviral copies/host DNA.
Discussion
Being classified as BLV ELISA-positive was positively associated with the probability of being pregnant; however, when qPCR-status or PVL-status were used to classify BLV-status, there was not a statistically significant association between the probability of being pregnant and BLV-status. However, an important limitation of our study that should temper this observation is that a majority of ELISA-negative cows were not tested for qPCR-status or PVL-status and our analysis assumed that ELISA-negative cows were both qPCR-negative and PVL-low; an assumption that should be tested further.
The other aspect of our primary objective was related but different in that it addressed the question of whether or not BLV-status was associated with the probability of becoming pregnant in the first 21 days of the breeding season. The same variables tested for their association with overall pregnancy probability were used to address this question and none of the 3 methods of classifying BLV-status were associated with the probability of being classified as pregnant in the first 21 days of the breeding season. Although the association of BLV ELISA-status on becoming pregnant early in the breeding season does not provide evidence for either a positive or negative association of BLV serostatus with pregnancy success early in the breeding season, the concern that only 13% (164/1,268) of ELISA-negative cows were tested for qPCR- and PVL-status and we assumed that untested ELISA-negative cows were qPCR-negative and PVL-low tempers the interpretation of the qPCR and PVL results.
The results of this cross-sectional study of 43 beef cow-calf herds found that although being classified as BLV ELISA-positive was associated with a higher probability of being pregnant, the other methods of BLV-status classification did not provide evidence of an association. Similarly, none of the methods to classify BLV-status were associated with the probability of being classified as becoming pregnant in the first 21 days of the breeding season. We conclude that this study did not find evidence that testing beef cows for BLV-status and removing test-positive cows will improve cowherd fertility as described by the probability of becoming pregnant during the breeding season or becoming pregnant during the first 21 days of the breeding season
Although not a primary objective of this study, our findings that only 54% of the ELISA-positive cows were classified as qPCR-positive; and that of the 164 ELISA-negative cows that were evaluated by qPCR, 2.4% were classified as qPCR-positive, raise important questions for future studies to address to better understand BLV and diagnostic testing methods for this virus. Potential questions for further study include: Do cattle that are ELISA-positive, but PCR-negative retain those test results over time—or do they either become ELISA-negative or PCR-positive eventually? Does PVL change (increase) over time in ELISA-positive cattle? Are ELISA-positive but PCR-negative or low-PVL cattle likely to transmit BLV, and if so, what is the transmission risk compared with cattle classified as high-PVL?
Acknowledgments
We would like to thank Drs. Greg Hanzlicek and Kathryn Reif for their help with the laboratory analysis of samples. We also would like to thank the KSVDL, CentralStar Laboratories, and Advanced Animal Diagnostics for their time and contributions to this project. A special thanks go to CentralStar’s Research Laboratory director Casey Drosha PhD. We would also like to thank Kristen Smith for her help with the statistical analysis of the data.
References
- 1.↑
Gross L. Is leukemia caused by a transmissible virus? A working hypothesis. Blood. 1954;9:557–573. doi:10.1182/blood.V9.6.557.557
- 2.↑
Olson C, Hoss HE, Miller JM, et al. Evidence of bovine C-type (leukemia) virus in dairy cattle. JAVMA. 1973;163:355–357.
- 3.
Ferrer JF, Bhatt DM, Marshak RR, et al. Further studies on the antigenic properties and distribution of the putative bovine leukemia virus. In: Comparative Leukemia Research. Basel, Switzerland: S. Karger; 1973;59–66.
- 5.↑
Hopkins SG, DiGiacomo RF. Natural transmission of bovine leukemia virus in dairy and beef cattle. Vet Clin N Am. 1997;13(1):107–128.
- 6.↑
USDA. Bovine leukosis virus (BLV) in US beef cattle. #N299.299. Fort Collins, Colo: USDA, APHIS, Veterinary Services, National Animal Health Monitoring System; 1999. Accessed March 7, 2022. www.aphis.usda.gov/animal_health/nahms/beefcowcalf/downloads/beef97/Beef97_is_BLV.pdf
- 7.↑
Benitez OJ, Norby B, Bartlett PC, et al. Impact of bovine leukemia virus infection on beef cow longevity. Prev Vet Med. 2020;181:105055. doi:10.1016/j.prevetmed.2020.105055
- 8.↑
Benitez OJ, Robert JN, Norby B, et al. Lack of bovine leukemia virus transmission during natural breeding of cattle. Theriogenology. 2019;126:187–190. doi:10.1016/j.theriogenology.2018.12.005
- 9.↑
Jimba M, Takeshima S, Murakami H, et al. BLV-CoCoMo-qPCR: a useful tool for evaluating bovine leukemia virus infection status. BMC Vet Res. 2012;8:167. doi:10.1186/1746-6148-8-167
- 10.
Mekata H, Sekiguchi S, Konnai S, et al. Horizontal transmission and phylogenetic analysis of bovine leukemia virus in two districts of Miyazaki, Japan. J Vet Med Sci. 2015;77:1115–1120. doi:10.1292/jvms.14-0624
- 11.↑
Ruggiero VJ, Norby B, Benitez OJ, et al. Controlling bovine leukemia virus in dairy herds by identifying and removing cows with the highest proviral load and lymphocyte counts. J Dairy Sci. 2019;102:9165–9175. doi:10.3168/jds.2018-16186
- 12.↑
Kobayashi T, Inagaki Y, Ohnuki N, et al. Increasing bovine leukemia virus (BLV) proviral load is a risk factor for progression of enzootic bovine leukosis: a prospective study in Japan. Prev Vet Med. 2020;178:104680. doi:10.1016/j.prevetmed.2019.04.009
- 14.↑
Lewin HA, Wu MC, Nolan TJ, et al. Peripheral B lymphocyte percentage as an indicator of subclinical progression of bovine leukemia virus infection. J Dairy Sci. 1988;71:2526–2534. doi:10.3168/jds.S0022-0302(88)79841-0
- 15.↑
Alvarez I, Gutiérrez G, Gammella M, et al. Evaluation of total white blood cell count as a marker for proviral load of bovine leukemia virus in dairy cattle from herds with a high seroprevalence of antibodies against bovine leukemia virus. Am J Vet Res. 2013;74:744–749 doi:10.2460/ajvr.74.5.744
- 16.↑
Spire MF. Breeding season evaluation of beef herds. In: Howard JF, ed. Current Veterinary Therapy. 2 ed. Philadelphia, PA: WB Saunders, 1986;808–811.
- 17.
Bretzlaff KA. Pictorial guide to bovine pregnancy diagnosis. Vet Med. 1987;82:295–304.
- 18.
Larson RL. Evaluating information obtained from pregnancy examination in beef herds. Vet Med. 1999;94:566–576.
- 19.↑
Larson RL, White BJ. Evaluating information obtained from diagnosis of pregnancy ≥ of beef herds. Vet Clin North Am Food Anim Pract. 2016;32:319–224. doi:10.1016/j.cvfa.2016.01.005
- 20.↑
Richards MW, Spitzer JC, Warner MB. Effect of varying levels of postpartum nutrition and body condition at calving on subsequent reproductive performance in beef cattle. J Anim Sci. 1986;62:300–306. doi:10.2527/jas1986.622300x
- 21.↑
Jimba M, Takeshima S, Matoba K, et al. BLV-CoCoMo-qPCR: quantitation of bovine leukemia virus proviral load using the CoCoMo algorithm. Retrovirology. 2010;7:91. doi:10.1186/1742-4690-7-91
- 22.↑
Wickham H, Fransois R, Hemy L, et al. dplyr: a grammar of data; 2020. https://github.com/tidyverse/dplyr
- 23.↑
Wickham H. ggplot2: elegant graphics for data analysis: manipulation. R package version 1.0.1. New York, NY: Springer-Verlag; 2016. Accessed July 1, 2022. https://ggplot2.tidyverse.org
