Adenoviral vectors are used widely as gene therapy and vaccine delivery systems. An adenovirus-shedding assay may be performed in clinical trials to monitor the safety of the vector and to investigate the potential relation between clinical symptoms and shed vector virus. This report describes the development and statistical performance of the shedding assay. Live adenovirus was recovered from throat swab and urine samples spiked with E1-deleted adenovirus type 5 vector expressing HIV-1 gag [Ad5HIVgag], in the presence or absence of wild-type adenovirus (WT Ad5). Samples were cultured in 293 and A549 cells, and the DNA extracted from virus culture was tested by polymerase chain reaction (PCR) for sequence identity. The results showed that the frequency of Ad5HIVgag infectivity in 293 cells by cytopathic effect (CPE) or an immunofluorescence assay (IFA) was concentration-dependent (53% for 10(2), 94% for 10(4), and 100% for 10(6) viral particles). WT Ad5 virus did not interfere with Ad5HIVgag. PCR amplisets could specifically amplify target sequences in the background of nonspecific DNA matrices and could distinguish Ad5HIVgag from wild-type adenoviruses. This assay may be used for clinical trials using adenovirus vectors as vehicles for vaccines.