Background. Epidemiologic studies suggest that human herpesvirus 8 (HHV-8) is sexually transmitted among men who have sex with men; however, the mode of transmission is unclear. Methods. To evaluate the patterns of shedding of HHV-8, we obtained mucosal-secretion samples from a cohort of HHV-8-seropositive men who had sex with men and had no clinical evidence of Kaposi's sarcoma. Quantitative polymerase-chain-reaction (PCR) assays, in situ PCR assays, and in situ RNA hybridization were used to identify potential sources of infectious HHV-8. Results. We detected HHV-8 in at least one mucosal sample from 30 of 50 men who were seropositive for HHV-8 (60 percent). Overall, HHV-8 was detected in 30 percent of oropharyngeal samples, as compared with 1 percent of anal and genital samples (P < 0.001). In 39 percent of the HHV-8-seropositive men, HHV-8 was detected in saliva on more than 35 percent of the consecutive days on which samples were obtained. The median log titer of HHV-8 from the oral cavity was approximately 2.5 times as high as the titer at all other sites. In situ hybridization studies indicated that HHV-8 DNA and messenger RNA were present in oral epithelial cells. Among 92 men who had sex with men and who were seronegative for the human immunodeficiency virus (HIV), a history of sex with a partner who had Kaposi's sarcoma, deep kissing with an HIV-positive partner, and the use of amyl nitrite capsules ('poppers') or inhaled nitrites were independent risk factors for infection with HHV-8. Conclusions. Oral exposure to infectious saliva is a potential risk factor for the acquisition of HHV-8 among men who have sex with men. Hence, currently recommended safer sex practices may not protect against HHV-8 infection. (C) 2000, Massachusetts Medical Society.