Although the detection of antibodies to a specific pathogen is used initially as the assay of choice, direct detection of human retroviruses is difficult. First, only a small fraction of cells are infected in the peripheral blood and lymphatic tissue may serve as a reservoir for infection. Second, infected cells may harbor only a small number of copies of the viral sequences. Third, a latent infection marked by transcriptional dormancy is often established thereby obviating the use of proteins or RNA to detect the viruses. Fourth, closely related but distinct members of the onco-and lenti-virus families may complicate specific detection of a particular virus. An additional hurdle is viral heterogeneity. HIV variants, for example, have been identified within and among individuals harboring this virus. Accordingly, sensitive and specific detection of the human retroviruses seemingly requires specific amplification of viral DNA sequences prior to detection. In this regard, an in vitro DNA amplification procedure using DNA polymerase and termed the polymerase chain reaction (PCR) initially applied to human genetic diseases has been successfully applied to human retroviruses. A PCR-based assay has demonstrated utility for detecting infection: (1) prior to the generation of detectable antibodies, (2) in individuals with ambiguous or indeterminate serological status, (3) for neonatal screening, (4) by a specific type or multiple viruses, and (5) in therapeutic trials to allow the monitoring of infected cell load and viremia. It is also unlikely that the viruses identified to date represent all of the retroviruses responsible for human disease. Lymphatic disorders, in general, and immunodeficiencies, in particular, merit closer scrutiny for a retroviral etiologic agent.(ABSTRACT TRUNCATED AT 250 WORDS)