The use of the polymerase chain reaction (PCR) to detect the presence of Epstein-Barr virus (EBV) DNA in oral mucosa in the absence of specific lesions gives rise to the problem of identifying the real viral replication sites. To verify whether the detection of EBV is due to salivary contamination or its true replicative capacity in oral mucosa, saliva samples and exfoliated cells from four different oral mucosa sites were taken from 40 renal transplant patients and 20 normal subjects for examination by PCR using two pairs of primers specific for the BamHI-L and BamHI-K genomic regions. EBV-specific sequences were detected in one or more of the oral mucosa samples from 29 transplant patients (72.5%) and six healthy controls (30%), and in the saliva samples of 16 transplant patients (40%) and three healthy controls (15%). A total of 89 oral mucosa smears from 29 transplant patients, and 13 from healthy subjects, were EBV-positive. The positive samples were also investigated by means of in situ hybridization in order to confirm the intracellular presence of the viral genome, and by means of immunofluorescence testing with monoclonal antibodies to assess the possible expression of viral antigens. Hybridization with the EBV-specific probe was observed in 40/89 and 2/13 samples, respectively. Latent antigens (with or without lytic antigens) were detected in only 23 of the 40 samples (collected from eight different transplant patients) that were positive by in situ hybridization. Our data show that EBV is more frequently present in the oral mucosa of immunodeficient patients (where it can efficiently replicate) than in normal subjects.