Abstract
Enzootic bovine leukosis (EBL) is B-cell lymphoma in cattle caused by bovine leukemia virus (BLV) infection. The incidence of EBL has been increasing since 1998 in Japan, resulting in significant economic losses for farms. The BLV genome integrates with the host genome as provirus, leading to sustainably infection. Although most of the BLV-infected cattle are aleukemic, some cattle cause persistent lymphocytosis (PL) and subsequently develop EBL. Recent reports suggest the association between the risk for the transmission of BLV and the developing EBL and the proviral load (PVL) in BLV-infected cattle, which cannot measure readily in the field. This study aims to build a statistical model for predicting PVL of BLV-infected asymptomatic or PL cattle based on data accessible in the field. Five negative binomial regression models with different linear predictors were built and compared for the predictability of PVL. Consequently, the model with two explanatory variables (age in months and logarithm of lymphocyte count) was selected as the best model. The model can be used in the field as a cost-beneficial supporting tool to estimate the risk of transmission of BLV and developing EBL in infected cattle.