Two issues concerning virus detection and identification of persistently infected (PI) cattle were analysed in the study: 1) interference by maternal antibodies and 2) discrimination between PI and transiently infected (TI) animals. Antigen ELISA and RT-PCR based methods were compared using serum samples from natural and experimental PI and TI calves. RT-PCR and realtime RT-PCR using primers within 5’UTR region were more sensitive in detecting PI animals than Erns and NS3 antigen capture ELISAs, and they were not influenced by the presence of colostral antibodies in serum or by bovine viral diarrhoea virus genotype. The serum samples with Ct values ≤ 29.10 (corresponding to 104.87 viral RNA copies/μL) identified PI animals with 100% probability, while all samples with Ct values > 32.06 (corresponding to viral RNA load below 104 copies/μL) indicated TI status. The samples with Ct values between 29.10 and 32.06 (17.2% of PI and 11.5% of TI) should be considered as PI suspect and retested.